mariadb/mysql-test/t/mix_innodb_myisam_binlog.test

296 lines
7 KiB
Text
Raw Normal View History

2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
# Check that binlog is ok when a transaction mixes updates to InnoDB and
# MyISAM.
# It would be nice to make this a replication test, but in 4.0 the
# slave is always with --skip-innodb in the testsuite. I (Guilhem) however
# did some tests manually on a slave; tables are replicated fine and
Write slave status field names using consistent style. (Initial caps for each word.) For example, instead of writing Until_condition, Until_Log_File, and Until_log_pos, write Until_Condition, Until_Log_File, and Until_Log_pos. mysql-test/r/rpl000015.result: Write slave status field names using consistent style. mysql-test/r/rpl_empty_master_crash.result: Write slave status field names using consistent style. mysql-test/r/rpl_error_ignored_table.result: Write slave status field names using consistent style. mysql-test/r/rpl_flush_log_loop.result: Write slave status field names using consistent style. mysql-test/r/rpl_loaddata.result: Write slave status field names using consistent style. mysql-test/r/rpl_log.result: Write slave status field names using consistent style. mysql-test/r/rpl_log_pos.result: Write slave status field names using consistent style. mysql-test/r/rpl_max_relay_size.result: Write slave status field names using consistent style. mysql-test/r/rpl_openssl.result: Write slave status field names using consistent style. mysql-test/r/rpl_redirect.result: Write slave status field names using consistent style. mysql-test/r/rpl_replicate_do.result: Write slave status field names using consistent style. mysql-test/r/rpl_reset_slave.result: Write slave status field names using consistent style. mysql-test/r/rpl_rotate_logs.result: Write slave status field names using consistent style. mysql-test/r/rpl_trunc_binlog.result: Write slave status field names using consistent style. mysql-test/r/rpl_until.result: Write slave status field names using consistent style. mysql-test/t/mix_innodb_myisam_binlog.test: Write slave status field names using consistent style. mysql-test/t/rpl_max_relay_size.test: Write slave status field names using consistent style. sql/slave.cc: Write slave status field names using consistent style.
2003-11-20 20:07:25 +01:00
# Exec_Master_Log_Pos advances as expected.
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
# Embedded server doesn't support binlogging
-- source include/not_embedded.inc
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
-- source include/have_innodb.inc
--disable_warnings
drop table if exists t1, t2;
--enable_warnings
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
connect (con1,localhost,root,,);
connect (con2,localhost,root,,);
connection con1;
WorkLog#1323 Deprecate the use of TYPE=... Preferred syntax is ENGINE= include/mysqld_error.h: New warning for deprecated syntax sql/lex.h: Introduce ENGINE keyword Sort order of symbols sql/share/czech/errmsg.txt: New warning for deprecated syntax sql/share/danish/errmsg.txt: New warning for deprecated syntax sql/share/dutch/errmsg.txt: New warning for deprecated syntax sql/share/english/errmsg.txt: New warning for deprecated syntax sql/share/estonian/errmsg.txt: New warning for deprecated syntax sql/share/french/errmsg.txt: New warning for deprecated syntax sql/share/german/errmsg.txt: New warning for deprecated syntax sql/share/greek/errmsg.txt: New warning for deprecated syntax sql/share/hungarian/errmsg.txt: New warning for deprecated syntax sql/share/italian/errmsg.txt: New warning for deprecated syntax sql/share/japanese/errmsg.txt: New warning for deprecated syntax sql/share/korean/errmsg.txt: New warning for deprecated syntax sql/share/norwegian-ny/errmsg.txt: New warning for deprecated syntax sql/share/norwegian/errmsg.txt: New warning for deprecated syntax sql/share/polish/errmsg.txt: New warning for deprecated syntax sql/share/portuguese/errmsg.txt: New warning for deprecated syntax sql/share/romanian/errmsg.txt: New warning for deprecated syntax sql/share/russian/errmsg.txt: New warning for deprecated syntax sql/share/serbian/errmsg.txt: New warning for deprecated syntax sql/share/slovak/errmsg.txt: New warning for deprecated syntax sql/share/spanish/errmsg.txt: New warning for deprecated syntax sql/share/swedish/errmsg.txt: New warning for deprecated syntax sql/share/ukrainian/errmsg.txt: New warning for deprecated syntax sql/sql_show.cc: Change TYPE= with ENGINE= sql/sql_yacc.yy: Introduce ENGINE keyword, Deprecate TYPE= syntax, Introduce SHOW ENGINE syntax, Deprecate SHOW INNODB/BDB syntax. mysql-test/r/alias.result: Change occurances of TYPE= to ENGINE= mysql-test/r/alter_table.result: Change occurances of TYPE= to ENGINE= mysql-test/r/auto_increment.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb-alter-table-1.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb-crash.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb-deadlock.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb_cache.result: Change occurances of TYPE= to ENGINE= mysql-test/r/case.result: Change occurances of TYPE= to ENGINE= mysql-test/r/cast.result: Change occurances of TYPE= to ENGINE= mysql-test/r/constraints.result: Change occurances of TYPE= to ENGINE= mysql-test/r/create.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_collate.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_latin1_de.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_many.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_mb.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_recoding.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_ucs.result: Change occurances of TYPE= to ENGINE= mysql-test/r/delete.result: Change occurances of TYPE= to ENGINE= mysql-test/r/distinct.result: Change occurances of TYPE= to ENGINE= mysql-test/r/fulltext.result: Change occurances of TYPE= to ENGINE= mysql-test/r/fulltext2.result: Change occurances of TYPE= to ENGINE= mysql-test/r/fulltext_distinct.result: Change occurances of TYPE= to ENGINE= mysql-test/r/fulltext_left_join.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_compress.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_date_add.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_group.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_if.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_str.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_system.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_test.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_time.result: Change occurances of TYPE= to ENGINE= mysql-test/r/gis-rtree.result: Change occurances of TYPE= to ENGINE= mysql-test/r/group_by.result: Change occurances of TYPE= to ENGINE= mysql-test/r/handler.result: Change occurances of TYPE= to ENGINE= mysql-test/r/heap.result: Change occurances of TYPE= to ENGINE= mysql-test/r/heap_auto_increment.result: Change occurances of TYPE= to ENGINE= mysql-test/r/heap_btree.result: Change occurances of TYPE= to ENGINE= mysql-test/r/heap_hash.result: Change occurances of TYPE= to ENGINE= mysql-test/r/help.result: Change occurances of TYPE= to ENGINE= mysql-test/r/innodb-deadlock.result: Change occurances of TYPE= to ENGINE= mysql-test/r/innodb.result: Change occurances of TYPE= to ENGINE= mysql-test/r/innodb_cache.result: Change occurances of TYPE= to ENGINE= mysql-test/r/innodb_handler.result: Change occurances of TYPE= to ENGINE= mysql-test/r/insert_select.result: Change occurances of TYPE= to ENGINE= mysql-test/r/isam.result: Change occurances of TYPE= to ENGINE= mysql-test/r/join.result: Change occurances of TYPE= to ENGINE= mysql-test/r/join_crash.result: Change occurances of TYPE= to ENGINE= mysql-test/r/join_outer.result: Change occurances of TYPE= to ENGINE= mysql-test/r/key.result: Change occurances of TYPE= to ENGINE= mysql-test/r/lock.result: Change occurances of TYPE= to ENGINE= mysql-test/r/lock_tables_lost_commit.result: Change occurances of TYPE= to ENGINE= mysql-test/r/merge.result: Change occurances of TYPE= to ENGINE= mysql-test/r/mix_innodb_myisam_binlog.result: Change occurances of TYPE= to ENGINE= mysql-test/r/multi_update.result: Change occurances of TYPE= to ENGINE= mysql-test/r/myisam.result: Change occurances of TYPE= to ENGINE= mysql-test/r/null.result: Change occurances of TYPE= to ENGINE= mysql-test/r/null_key.result: Change occurances of TYPE= to ENGINE= mysql-test/r/order_by.result: Change occurances of TYPE= to ENGINE= mysql-test/r/query_cache.result: Change occurances of TYPE= to ENGINE= mysql-test/r/range.result: Change occurances of TYPE= to ENGINE= mysql-test/r/repair_part1.result: Change occurances of TYPE= to ENGINE= mysql-test/r/replace.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rollback.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rpl000006.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rpl_flush_tables.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rpl_insert_id.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rpl_relayrotate.result: Change occurances of TYPE= to ENGINE= mysql-test/r/select.result: Change occurances of TYPE= to ENGINE= mysql-test/r/select_found.result: Change occurances of TYPE= to ENGINE= mysql-test/r/show_check.result: Change occurances of TYPE= to ENGINE= mysql-test/r/sql_mode.result: Change occurances of TYPE= to ENGINE= mysql-test/r/status.result: Change occurances of TYPE= to ENGINE= mysql-test/r/subselect.result: Change occurances of TYPE= to ENGINE= mysql-test/r/subselect2.result: Change occurances of TYPE= to ENGINE= mysql-test/r/subselect_innodb.result: Change occurances of TYPE= to ENGINE= mysql-test/r/symlink.result: Change occurances of TYPE= to ENGINE= mysql-test/r/temp_table.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_blob.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_datetime.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_enum.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_nchar.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_set.result: Change occurances of TYPE= to ENGINE= mysql-test/r/union.result: Change occurances of TYPE= to ENGINE= mysql-test/r/update.result: Change occurances of TYPE= to ENGINE= mysql-test/r/warnings.result: Change occurances of TYPE= to ENGINE= mysql-test/t/alias.test: Change occurances of TYPE= to ENGINE= mysql-test/t/alter_table.test: Change occurances of TYPE= to ENGINE= mysql-test/t/auto_increment.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb-alter-table-1.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb-crash.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb-deadlock.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb_cache.test: Change occurances of TYPE= to ENGINE= mysql-test/t/create.test: Change occurances of TYPE= to ENGINE= mysql-test/t/ctype_ucs.test: Change occurances of TYPE= to ENGINE= mysql-test/t/delete.test: Change occurances of TYPE= to ENGINE= mysql-test/t/distinct.test: Change occurances of TYPE= to ENGINE= mysql-test/t/fulltext.test: Change occurances of TYPE= to ENGINE= mysql-test/t/fulltext2.test: Change occurances of TYPE= to ENGINE= mysql-test/t/fulltext_distinct.test: Change occurances of TYPE= to ENGINE= mysql-test/t/fulltext_left_join.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_compress.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_date_add.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_group.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_if.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_str.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_test.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_time.test: Change occurances of TYPE= to ENGINE= mysql-test/t/gis-rtree.test: Change occurances of TYPE= to ENGINE= mysql-test/t/group_by.test: Change occurances of TYPE= to ENGINE= mysql-test/t/handler.test: Change occurances of TYPE= to ENGINE= mysql-test/t/heap.test: Change occurances of TYPE= to ENGINE= mysql-test/t/heap_auto_increment.test: Change occurances of TYPE= to ENGINE= mysql-test/t/heap_btree.test: Change occurances of TYPE= to ENGINE= mysql-test/t/heap_hash.test: Change occurances of TYPE= to ENGINE= mysql-test/t/help.test: Change occurances of TYPE= to ENGINE= mysql-test/t/innodb-deadlock.test: Change occurances of TYPE= to ENGINE= mysql-test/t/innodb.test: Change occurances of TYPE= to ENGINE= mysql-test/t/innodb_cache.test: Change occurances of TYPE= to ENGINE= mysql-test/t/innodb_handler.test: Change occurances of TYPE= to ENGINE= mysql-test/t/insert_select.test: Change occurances of TYPE= to ENGINE= mysql-test/t/isam.test: Change occurances of TYPE= to ENGINE= mysql-test/t/join.test: Change occurances of TYPE= to ENGINE= mysql-test/t/join_crash.test: Change occurances of TYPE= to ENGINE= mysql-test/t/join_outer.test: Change occurances of TYPE= to ENGINE= mysql-test/t/key.test: Change occurances of TYPE= to ENGINE= mysql-test/t/lock.test: Change occurances of TYPE= to ENGINE= mysql-test/t/lock_tables_lost_commit.test: Change occurances of TYPE= to ENGINE= mysql-test/t/merge.test: Change occurances of TYPE= to ENGINE= mysql-test/t/mix_innodb_myisam_binlog.test: Change occurances of TYPE= to ENGINE= mysql-test/t/multi_update.test: Change occurances of TYPE= to ENGINE= mysql-test/t/myisam.test: Change occurances of TYPE= to ENGINE= mysql-test/t/null.test: Change occurances of TYPE= to ENGINE= mysql-test/t/null_key.test: Change occurances of TYPE= to ENGINE= mysql-test/t/order_by.test: Change occurances of TYPE= to ENGINE= mysql-test/t/outfile.test: Change occurances of TYPE= to ENGINE= mysql-test/t/query_cache.test: Change occurances of TYPE= to ENGINE= mysql-test/t/query_cache_merge.test: Change occurances of TYPE= to ENGINE= mysql-test/t/range.test: Change occurances of TYPE= to ENGINE= mysql-test/t/repair_part1.test: Change occurances of TYPE= to ENGINE= mysql-test/t/replace.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rollback.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rpl000006.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rpl_flush_tables.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rpl_insert_id.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rpl_relayrotate.test: Change occurances of TYPE= to ENGINE= mysql-test/t/select.test: Change occurances of TYPE= to ENGINE= mysql-test/t/select_found.test: Change occurances of TYPE= to ENGINE= mysql-test/t/show_check.test: Change occurances of TYPE= to ENGINE= mysql-test/t/sql_mode.test: Change occurances of TYPE= to ENGINE= mysql-test/t/status.test: Change occurances of TYPE= to ENGINE= mysql-test/t/subselect.test: Change occurances of TYPE= to ENGINE= mysql-test/t/subselect2.test: Change occurances of TYPE= to ENGINE= mysql-test/t/subselect_innodb.test: Change occurances of TYPE= to ENGINE= mysql-test/t/symlink.test: Change occurances of TYPE= to ENGINE= mysql-test/t/temp_table.test: Change occurances of TYPE= to ENGINE= mysql-test/t/type_datetime.test: Change occurances of TYPE= to ENGINE= mysql-test/t/type_set.test: Change occurances of TYPE= to ENGINE= mysql-test/t/union.test: Change occurances of TYPE= to ENGINE= mysql-test/t/update.test: Change occurances of TYPE= to ENGINE= mysql-test/t/warnings.test: Change occurances of TYPE= to ENGINE= New test for deprecated syntax
2003-12-10 05:31:42 +01:00
create table t1 (a int) engine=innodb;
create table t2 (a int) engine=myisam;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
begin;
insert into t1 values(1);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
commit;
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=15" "xid=8"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
begin;
insert into t1 values(2);
insert into t2 select * from t1;
# should say some changes to non-transact1onal tables couldn't be rolled back
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
rollback;
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
begin;
insert into t1 values(3);
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
savepoint my_savepoint;
insert into t1 values(4);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
rollback to savepoint my_savepoint;
commit;
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=48" "xid=25"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
begin;
insert into t1 values(5);
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
savepoint my_savepoint;
insert into t1 values(6);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
rollback to savepoint my_savepoint;
insert into t1 values(7);
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
commit;
select a from t1 order by a; # check that savepoints work :)
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=70" "xid=37"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
# and when ROLLBACK is not explicit?
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
select get_lock("a",10);
begin;
insert into t1 values(8);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
disconnect con1;
connection con2;
# We want to SHOW BINLOG EVENTS, to know what was logged. But there is no
# guarantee that logging of the terminated con1 has been done yet (it may not
# even be started, so con1 may have not even attempted to lock the binlog yet;
# so SHOW BINLOG EVENTS may come before con1 does the loggin. To be sure that
# logging has been done, we use a user lock.
select get_lock("a",10);
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
# and when not in a transact1on?
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
insert into t1 values(9);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=119" "xid=60"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
# Check that when the query updat1ng the MyISAM table is the first in the
# transaction, we log it immediately.
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
insert into t1 values(10); # first make t1 non-empty
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
begin;
insert into t2 select * from t1;
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=133" "xid=66"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
insert into t1 values(11);
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
commit;
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=133" "xid=66" "xid=136" "xid=68"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
# Check that things work like before this BEGIN/ROLLBACK code was added,
# when t2 is INNODB
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
WorkLog#1323 Deprecate the use of TYPE=... Preferred syntax is ENGINE= include/mysqld_error.h: New warning for deprecated syntax sql/lex.h: Introduce ENGINE keyword Sort order of symbols sql/share/czech/errmsg.txt: New warning for deprecated syntax sql/share/danish/errmsg.txt: New warning for deprecated syntax sql/share/dutch/errmsg.txt: New warning for deprecated syntax sql/share/english/errmsg.txt: New warning for deprecated syntax sql/share/estonian/errmsg.txt: New warning for deprecated syntax sql/share/french/errmsg.txt: New warning for deprecated syntax sql/share/german/errmsg.txt: New warning for deprecated syntax sql/share/greek/errmsg.txt: New warning for deprecated syntax sql/share/hungarian/errmsg.txt: New warning for deprecated syntax sql/share/italian/errmsg.txt: New warning for deprecated syntax sql/share/japanese/errmsg.txt: New warning for deprecated syntax sql/share/korean/errmsg.txt: New warning for deprecated syntax sql/share/norwegian-ny/errmsg.txt: New warning for deprecated syntax sql/share/norwegian/errmsg.txt: New warning for deprecated syntax sql/share/polish/errmsg.txt: New warning for deprecated syntax sql/share/portuguese/errmsg.txt: New warning for deprecated syntax sql/share/romanian/errmsg.txt: New warning for deprecated syntax sql/share/russian/errmsg.txt: New warning for deprecated syntax sql/share/serbian/errmsg.txt: New warning for deprecated syntax sql/share/slovak/errmsg.txt: New warning for deprecated syntax sql/share/spanish/errmsg.txt: New warning for deprecated syntax sql/share/swedish/errmsg.txt: New warning for deprecated syntax sql/share/ukrainian/errmsg.txt: New warning for deprecated syntax sql/sql_show.cc: Change TYPE= with ENGINE= sql/sql_yacc.yy: Introduce ENGINE keyword, Deprecate TYPE= syntax, Introduce SHOW ENGINE syntax, Deprecate SHOW INNODB/BDB syntax. mysql-test/r/alias.result: Change occurances of TYPE= to ENGINE= mysql-test/r/alter_table.result: Change occurances of TYPE= to ENGINE= mysql-test/r/auto_increment.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb-alter-table-1.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb-crash.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb-deadlock.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb.result: Change occurances of TYPE= to ENGINE= mysql-test/r/bdb_cache.result: Change occurances of TYPE= to ENGINE= mysql-test/r/case.result: Change occurances of TYPE= to ENGINE= mysql-test/r/cast.result: Change occurances of TYPE= to ENGINE= mysql-test/r/constraints.result: Change occurances of TYPE= to ENGINE= mysql-test/r/create.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_collate.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_latin1_de.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_many.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_mb.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_recoding.result: Change occurances of TYPE= to ENGINE= mysql-test/r/ctype_ucs.result: Change occurances of TYPE= to ENGINE= mysql-test/r/delete.result: Change occurances of TYPE= to ENGINE= mysql-test/r/distinct.result: Change occurances of TYPE= to ENGINE= mysql-test/r/fulltext.result: Change occurances of TYPE= to ENGINE= mysql-test/r/fulltext2.result: Change occurances of TYPE= to ENGINE= mysql-test/r/fulltext_distinct.result: Change occurances of TYPE= to ENGINE= mysql-test/r/fulltext_left_join.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_compress.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_date_add.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_group.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_if.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_str.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_system.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_test.result: Change occurances of TYPE= to ENGINE= mysql-test/r/func_time.result: Change occurances of TYPE= to ENGINE= mysql-test/r/gis-rtree.result: Change occurances of TYPE= to ENGINE= mysql-test/r/group_by.result: Change occurances of TYPE= to ENGINE= mysql-test/r/handler.result: Change occurances of TYPE= to ENGINE= mysql-test/r/heap.result: Change occurances of TYPE= to ENGINE= mysql-test/r/heap_auto_increment.result: Change occurances of TYPE= to ENGINE= mysql-test/r/heap_btree.result: Change occurances of TYPE= to ENGINE= mysql-test/r/heap_hash.result: Change occurances of TYPE= to ENGINE= mysql-test/r/help.result: Change occurances of TYPE= to ENGINE= mysql-test/r/innodb-deadlock.result: Change occurances of TYPE= to ENGINE= mysql-test/r/innodb.result: Change occurances of TYPE= to ENGINE= mysql-test/r/innodb_cache.result: Change occurances of TYPE= to ENGINE= mysql-test/r/innodb_handler.result: Change occurances of TYPE= to ENGINE= mysql-test/r/insert_select.result: Change occurances of TYPE= to ENGINE= mysql-test/r/isam.result: Change occurances of TYPE= to ENGINE= mysql-test/r/join.result: Change occurances of TYPE= to ENGINE= mysql-test/r/join_crash.result: Change occurances of TYPE= to ENGINE= mysql-test/r/join_outer.result: Change occurances of TYPE= to ENGINE= mysql-test/r/key.result: Change occurances of TYPE= to ENGINE= mysql-test/r/lock.result: Change occurances of TYPE= to ENGINE= mysql-test/r/lock_tables_lost_commit.result: Change occurances of TYPE= to ENGINE= mysql-test/r/merge.result: Change occurances of TYPE= to ENGINE= mysql-test/r/mix_innodb_myisam_binlog.result: Change occurances of TYPE= to ENGINE= mysql-test/r/multi_update.result: Change occurances of TYPE= to ENGINE= mysql-test/r/myisam.result: Change occurances of TYPE= to ENGINE= mysql-test/r/null.result: Change occurances of TYPE= to ENGINE= mysql-test/r/null_key.result: Change occurances of TYPE= to ENGINE= mysql-test/r/order_by.result: Change occurances of TYPE= to ENGINE= mysql-test/r/query_cache.result: Change occurances of TYPE= to ENGINE= mysql-test/r/range.result: Change occurances of TYPE= to ENGINE= mysql-test/r/repair_part1.result: Change occurances of TYPE= to ENGINE= mysql-test/r/replace.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rollback.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rpl000006.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rpl_flush_tables.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rpl_insert_id.result: Change occurances of TYPE= to ENGINE= mysql-test/r/rpl_relayrotate.result: Change occurances of TYPE= to ENGINE= mysql-test/r/select.result: Change occurances of TYPE= to ENGINE= mysql-test/r/select_found.result: Change occurances of TYPE= to ENGINE= mysql-test/r/show_check.result: Change occurances of TYPE= to ENGINE= mysql-test/r/sql_mode.result: Change occurances of TYPE= to ENGINE= mysql-test/r/status.result: Change occurances of TYPE= to ENGINE= mysql-test/r/subselect.result: Change occurances of TYPE= to ENGINE= mysql-test/r/subselect2.result: Change occurances of TYPE= to ENGINE= mysql-test/r/subselect_innodb.result: Change occurances of TYPE= to ENGINE= mysql-test/r/symlink.result: Change occurances of TYPE= to ENGINE= mysql-test/r/temp_table.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_blob.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_datetime.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_enum.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_nchar.result: Change occurances of TYPE= to ENGINE= mysql-test/r/type_set.result: Change occurances of TYPE= to ENGINE= mysql-test/r/union.result: Change occurances of TYPE= to ENGINE= mysql-test/r/update.result: Change occurances of TYPE= to ENGINE= mysql-test/r/warnings.result: Change occurances of TYPE= to ENGINE= mysql-test/t/alias.test: Change occurances of TYPE= to ENGINE= mysql-test/t/alter_table.test: Change occurances of TYPE= to ENGINE= mysql-test/t/auto_increment.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb-alter-table-1.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb-crash.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb-deadlock.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb.test: Change occurances of TYPE= to ENGINE= mysql-test/t/bdb_cache.test: Change occurances of TYPE= to ENGINE= mysql-test/t/create.test: Change occurances of TYPE= to ENGINE= mysql-test/t/ctype_ucs.test: Change occurances of TYPE= to ENGINE= mysql-test/t/delete.test: Change occurances of TYPE= to ENGINE= mysql-test/t/distinct.test: Change occurances of TYPE= to ENGINE= mysql-test/t/fulltext.test: Change occurances of TYPE= to ENGINE= mysql-test/t/fulltext2.test: Change occurances of TYPE= to ENGINE= mysql-test/t/fulltext_distinct.test: Change occurances of TYPE= to ENGINE= mysql-test/t/fulltext_left_join.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_compress.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_date_add.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_group.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_if.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_str.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_test.test: Change occurances of TYPE= to ENGINE= mysql-test/t/func_time.test: Change occurances of TYPE= to ENGINE= mysql-test/t/gis-rtree.test: Change occurances of TYPE= to ENGINE= mysql-test/t/group_by.test: Change occurances of TYPE= to ENGINE= mysql-test/t/handler.test: Change occurances of TYPE= to ENGINE= mysql-test/t/heap.test: Change occurances of TYPE= to ENGINE= mysql-test/t/heap_auto_increment.test: Change occurances of TYPE= to ENGINE= mysql-test/t/heap_btree.test: Change occurances of TYPE= to ENGINE= mysql-test/t/heap_hash.test: Change occurances of TYPE= to ENGINE= mysql-test/t/help.test: Change occurances of TYPE= to ENGINE= mysql-test/t/innodb-deadlock.test: Change occurances of TYPE= to ENGINE= mysql-test/t/innodb.test: Change occurances of TYPE= to ENGINE= mysql-test/t/innodb_cache.test: Change occurances of TYPE= to ENGINE= mysql-test/t/innodb_handler.test: Change occurances of TYPE= to ENGINE= mysql-test/t/insert_select.test: Change occurances of TYPE= to ENGINE= mysql-test/t/isam.test: Change occurances of TYPE= to ENGINE= mysql-test/t/join.test: Change occurances of TYPE= to ENGINE= mysql-test/t/join_crash.test: Change occurances of TYPE= to ENGINE= mysql-test/t/join_outer.test: Change occurances of TYPE= to ENGINE= mysql-test/t/key.test: Change occurances of TYPE= to ENGINE= mysql-test/t/lock.test: Change occurances of TYPE= to ENGINE= mysql-test/t/lock_tables_lost_commit.test: Change occurances of TYPE= to ENGINE= mysql-test/t/merge.test: Change occurances of TYPE= to ENGINE= mysql-test/t/mix_innodb_myisam_binlog.test: Change occurances of TYPE= to ENGINE= mysql-test/t/multi_update.test: Change occurances of TYPE= to ENGINE= mysql-test/t/myisam.test: Change occurances of TYPE= to ENGINE= mysql-test/t/null.test: Change occurances of TYPE= to ENGINE= mysql-test/t/null_key.test: Change occurances of TYPE= to ENGINE= mysql-test/t/order_by.test: Change occurances of TYPE= to ENGINE= mysql-test/t/outfile.test: Change occurances of TYPE= to ENGINE= mysql-test/t/query_cache.test: Change occurances of TYPE= to ENGINE= mysql-test/t/query_cache_merge.test: Change occurances of TYPE= to ENGINE= mysql-test/t/range.test: Change occurances of TYPE= to ENGINE= mysql-test/t/repair_part1.test: Change occurances of TYPE= to ENGINE= mysql-test/t/replace.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rollback.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rpl000006.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rpl_flush_tables.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rpl_insert_id.test: Change occurances of TYPE= to ENGINE= mysql-test/t/rpl_relayrotate.test: Change occurances of TYPE= to ENGINE= mysql-test/t/select.test: Change occurances of TYPE= to ENGINE= mysql-test/t/select_found.test: Change occurances of TYPE= to ENGINE= mysql-test/t/show_check.test: Change occurances of TYPE= to ENGINE= mysql-test/t/sql_mode.test: Change occurances of TYPE= to ENGINE= mysql-test/t/status.test: Change occurances of TYPE= to ENGINE= mysql-test/t/subselect.test: Change occurances of TYPE= to ENGINE= mysql-test/t/subselect2.test: Change occurances of TYPE= to ENGINE= mysql-test/t/subselect_innodb.test: Change occurances of TYPE= to ENGINE= mysql-test/t/symlink.test: Change occurances of TYPE= to ENGINE= mysql-test/t/temp_table.test: Change occurances of TYPE= to ENGINE= mysql-test/t/type_datetime.test: Change occurances of TYPE= to ENGINE= mysql-test/t/type_set.test: Change occurances of TYPE= to ENGINE= mysql-test/t/union.test: Change occurances of TYPE= to ENGINE= mysql-test/t/update.test: Change occurances of TYPE= to ENGINE= mysql-test/t/warnings.test: Change occurances of TYPE= to ENGINE= New test for deprecated syntax
2003-12-10 05:31:42 +01:00
alter table t2 engine=INNODB;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
begin;
insert into t1 values(12);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
commit;
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=155" "xid=78"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
begin;
insert into t1 values(13);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
rollback;
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
begin;
insert into t1 values(14);
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
savepoint my_savepoint;
insert into t1 values(15);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
rollback to savepoint my_savepoint;
commit;
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=187" "xid=94"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
delete from t1;
delete from t2;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
reset master;
begin;
insert into t1 values(16);
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
savepoint my_savepoint;
insert into t1 values(17);
insert into t2 select * from t1;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
rollback to savepoint my_savepoint;
insert into t1 values(18);
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
commit;
select a from t1 order by a; # check that savepoints work :)
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
Added options --auto-increment-increment and --auto-increment-offset. This allows one to setup a master <-> master replication with non conflicting auto-increment series. Cleaned up binary log code to make it easyer to add new state variables. Added simpler 'upper level' logic for artificial events (events that should not cause cleanups on slave). Simplified binary log handling. Changed how auto_increment works together with to SET INSERT_ID=# to make it more predictable: Now the inserted rows in a multi-row statement are set independent of the existing rows in the table. (Before only InnoDB did this correctly) mysql-test/r/mix_innodb_myisam_binlog.result: Disable End_log_pos column from 'show binlog events' as this is now different from before mysql-test/t/mix_innodb_myisam_binlog.test: Disable End_log_pos column from 'show binlog events' as this is now different from before sql/ha_berkeley.cc: Changed prototype for get_auto_increment() sql/ha_berkeley.h: Changed prototype for get_auto_increment() sql/ha_heap.cc: Changed prototype for get_auto_increment() sql/ha_heap.h: Changed prototype for get_auto_increment() sql/ha_innodb.cc: Change how auto-increment is calculated. Now the auto-increment logic is done in 'update_auto_increment()' to ensure that all handlers has the same auto-increment usage sql/ha_innodb.h: Changed prototype for get_auto_increment() sql/ha_myisam.cc: Changed prototype for get_auto_increment() sql/ha_myisam.h: Changed prototype for get_auto_increment() sql/ha_ndbcluster.cc: Changed prototype for get_auto_increment() sql/ha_ndbcluster.h: Changed prototype for get_auto_increment() sql/handler.cc: Remove some usage of current_thd Changed how auto_increment works with SET INSERT_ID to make it more predictable (Now we should generate same auto-increment serie on a slave, even if the table has rows that was not on the master. Use auto_increment_increment and auto_increment_offset sql/handler.h: Changed prototype for get_auto_increment() sql/log.cc: Remove usage of 'set_log_pos()' to make code simpler. (Now log_pos is set in write_header()) Use 'data_written' instead of 'get_event_len()' to calculate how much data was written in the log sql/log_event.cc: Simple optimizations. Remove cached_event_len (not used variable) Made comments fit into 79 chars Removed Log_event::set_log_pos(). Now we calculate log_pos in write_header(). Renamed write_data() to write() as the original write() function was not needed anymore. Call writing of event header from event::write() functions. This made it easier to calculate the length of an event. Simplified 'write_header' and remove 'switches' from it. Changed all write() functions to return 'bool'. (The previous return values where not consistent) Store auto_increment_increment and auto_increment_offset in binary log Simplified how Query_log_event's where written and read. Now it's much easier to add now status variables for a query event to the binary log. Removed some old MySQL 4.x code to make it easier to grep for functions used in 5.0 sql/log_event.h: Changed return type of write() functions to bool. (Before we returned -1 or 1 for errors) write_data() -> write() Added 'data_written' member to make it easier to get length of written event. Removed 'cached_event_len' and 'get_event_len()' Added usage of auto_increment_increment and auto_increment_offset Added 'artifical_event' to Start_log_event_v3, to hide logic that we in the binary log use log_pos=0 as a flag for an artifical event. sql/mysqld.cc: Added options --auto-increment-increment and --auto-increment-offset sql/set_var.cc: Added variables auto_increment_increment and auto_increment_offset sql/slave.cc: Changed errors -> warnings & information (in error log) sql/sql_class.cc: Added THD::cleanup_after_query(). This makes some code simpler and allows us to clean up 'next_insert_id' after query sql/sql_class.h: Added new auto_increment_xxx variables Moved some functions/variables in THD class sql/sql_help.cc: Removed compiler warning sql/sql_insert.cc: Call 'restore_auto_increment()' if row was not inserted. This makes it easier for handler to reuse the last generated auto-incrment value that was not used (for example in case of duplicate key) sql/sql_parse.cc: Use cleanup_after_query() sql/sql_prepare.cc: Use cleanup_after_query() sql/sql_table.cc: R
2004-09-15 21:10:31 +02:00
--replace_column 5 #
--replace_result "xid=208" "xid=105"
WL#874 "Extended LOAD DATA". Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; This patch also implements new way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event client/mysqlbinlog.cc: Added support of two new binary log events Begin_load_query_log_event and Execute_load_query_log_Event which are used to replicate LOAD DATA INFILE. mysql-test/r/ctype_ucs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/insert_select.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/loaddata.result: Added tests for new LOAD DATA features. mysql-test/r/mix_innodb_myisam_binlog.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/mysqlbinlog.result: New approach for binlogging of LOAD DATA statement. Now we store it as usual query and rewrite part in which file is specified when needed. So now mysqlbinlog output for LOAD DATA much more closer to its initial form. Updated test'd results accordingly. mysql-test/r/mysqldump.result: Made test more robust to other tests failures. mysql-test/r/rpl000015.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_change_master.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results. mysql-test/r/rpl_charset.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_deadlock.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly mysql-test/r/rpl_error_ignored_table.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_flush_log_loop.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_flush_tables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_loaddata.result: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Updated test's results wwith new binlog positions. mysql-test/r/rpl_loaddata_rule_m.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/r/rpl_loaddata_rule_s.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_loaddatalocal.result: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/r/rpl_log.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/r/rpl_log_pos.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_max_relay_size.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_multi_query.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_relayrotate.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_replicate_do.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_reset_slave.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_rotate_logs.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id1.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_server_id2.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly. mysql-test/r/rpl_temporary.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_timezone.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/rpl_until.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results accordingly and tweaked test a bit to bring it back to good shape. mysql-test/r/rpl_user_variables.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/r/user_var.result: Addition of two new types of binary log events shifted binlog positions. Updated test's results and made it more robust for future similar changes. mysql-test/t/ctype_ucs.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/insert_select.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/loaddata.test: Added test cases for new LOAD DATA functionality. mysql-test/t/mix_innodb_myisam_binlog.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqlbinlog.test: New way of replicating LOAD DATA local. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Thus we need new binlog positions for LOAD DATA events. mysql-test/t/mysqlbinlog2.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/mysqldump.test: Made test more robust for failures of other tests. mysql-test/t/rpl_charset.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_deadlock.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_error_ignored_table.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/rpl_flush_tables.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_loaddata.test: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event... Apropritely updated comments in test. mysql-test/t/rpl_loaddata_rule_m.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. Since now LOAD DATA is replicated much in the same way as usual query --binlog_do/ignore_db work for it inthe same way as for usual queries. mysql-test/t/rpl_loaddata_rule_s.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_loaddatalocal.test: Added nice test for case when it is important that LOAD DATA LOCAL ignores duplicates. mysql-test/t/rpl_log.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly (don't dare to get rid from binlog positions completely since it seems that this test uses them). mysql-test/t/rpl_log_pos.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_multi_query.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly. mysql-test/t/rpl_temporary.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_timezone.test: Addition of two new types of binary log events shifted binlog positions. Made test more robust for future similar changes. mysql-test/t/rpl_until.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and tweaked it a bit to bring it back to good shape. mysql-test/t/rpl_user_variables.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. mysql-test/t/user_var.test: Addition of two new types of binary log events shifted binlog positions. Updated test accordingly and made it more robust for future similar changes. sql/item_func.cc: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. Moved code from Item_func_set_user_var::update_hash() function to separate static function to be able to reuse it in this new class. sql/item_func.h: Added Item_user_var_as_out_param class that represents user variable which used as out parameter in LOAD DATA. sql/log_event.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/log_event.h: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/mysql_priv.h: Now mysql_load() has two more arguments. They are needed to pass list of columns and corresponding expressions from new LOAD DATA's SET clause. sql/share/errmsg.txt: Added new error message which is used to forbid loading of data from fixed length rows to variables. sql/sql_lex.h: Added LEX::fname_start/fname_end members. They are pointers to part of LOAD DATA statement which should be rewritten during replication (file name + little extra). sql/sql_load.cc: Added support for extended LOAD DATA. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. Updated mysql_load()/read_fixed_length()/read_sep_field() to support this functionality (now they can read data from file to both columns and variables and assign do calculations and assignments specified in SET clause). We also use new approach for LOAD DATA binlogging/replication. sql/sql_parse.cc: mysql_execute_command(): Since now we have SET clause in LOAD DATA we should also check permissions for tables used in its expressions. Also mysql_load() has two more arguments to pass information about this clause. sql/sql_repl.cc: New way of replicating LOAD DATA. Now we do it similarly to other queries. We store LOAD DATA query in new Execute_load_query event (which is last in the sequence of events representing LOAD DATA). When we are executing this event we simply rewrite part of query which holds name of file (we use name of temporary file) and then execute it as usual query. In the beggining of this sequence we use Begin_load_query event which is almost identical to Append_file event. sql/sql_repl.h: struct st_load_file_info: Removed memebers which are no longer needed for LOAD DATA binnlogging. sql/sql_yacc.yy: Added support for extended LOAD DATA syntax. Now one can use user variables as target for data loaded from file (besides table's columns). Also LOAD DATA got new SET-clause in which one can specify values for table columns as expressions. For example the following is possible: LOAD DATA INFILE 'words.dat' INTO TABLE t1 (a, @b) SET c = @b + 1; Also now we save pointers to the beginning and to the end of part of LOAD DATA statement which should be rewritten during replication.
2005-03-16 02:32:47 +01:00
show binlog events from 98;
2 minor edits, plus fix for BUG#1113 "INSERT into non-trans table SELECT ; ROLLBACK" does not send warning" and fix for BUG#873 "In transaction, INSERT to non-trans table is written too early to binlog". Now we don't always write the non-trans update immediately to the binlog; if there is something in the binlog cache we write it to the binlog cache (because the non-trans update could depend on a trans table which was modified earlier in the transaction); then in case of ROLLBACK, we write the binlog cache to the binlog, wrapped with BEGIN/ROLLBACK. This guarantees that the slave does the same updates. For ROLLBACK TO SAVEPOINT: when we execute a SAVEPOINT command we write it to the binlog cache. At ROLLBACK TO SAVEPOINT, if some non-trans table was updated, we write ROLLBACK TO SAVEPOINT to the binlog cache; when the transaction terminates (COMMIT/ROLLBACK), the binlog cache will be flushed to the binlog (because of the non-trans update) so we'll have SAVEPOINT and ROLLBACK TO SAVEPOINT in the binlog. Apart from this rare case of updates of mixed table types in transaction, the usual way is still clear the binlog cache at ROLLBACK, or chop it at ROLLBACK TO SAVEPOINT (meaning the SAVEPOINT command is also chopped, which is fine). Note that BUG#873 encompasses subbugs 1) and 2) of BUG#333 "3 binlogging bugs when doing INSERT with mixed InnoDB/MyISAM". client/mysqldump.c: Minor edit: one CHANGE MASTER with 2 arguments instead of 2 CHANGE MASTER with one argument each. mysql-test/r/rpl_loaddata.result: result update mysql-test/t/rpl_loaddata.test: minor edit: simplifying the test. sql/handler.cc: Fix for BUG#873. See comments in code, and the description of the changeset. sql/log.cc: * Previously, if a query updated a non-transactional table we wrote it immediately to the real binlog. This causes a bug when the update is done inside a transaction and uses the content of an updated transactional table (because this makes a wrong order of queries in the binlog). So if the binlog cache is not empty, we write the query to the binlog cache; otherwise we can write it to the binlog. * Previously, when we flushed the binlog cache to the binlog, we wrapped it with BEGIN/COMMIT. Now it's also possible to wrap it with BEGIN/ROLLBACK, to handle transactions which update both transactional and non-transactional tables. sql/log_event.cc: The slave thread can leave a transaction if COMMIT or if ROLLBACK. sql/sql_class.h: prototype sql/sql_insert.cc: Fix for BUG#1113: this was because the INSERT SELECT code did not set OPTION_STATUS_NO_TRANS_UPDATE. sql/sql_parse.cc: Don't send ER_WARNING_NOT_COMPLETE_ROLLBACK if this is the SQL slave thread (see comments).
2003-08-22 15:39:24 +02:00
# Test for BUG#5714, where a MyISAM update in the transaction used to
# release row-level locks in InnoDB
connect (con3,localhost,root,,);
connection con3;
delete from t1;
delete from t2;
--disable_warnings
alter table t2 type=MyISAM;
--enable_warnings
insert into t1 values (1);
begin;
select * from t1 for update;
connection con2;
select (@before:=unix_timestamp())*0; # always give repeatable output
begin;
send select * from t1 for update;
connection con3;
insert into t2 values (20);
connection con2;
--error 1205
reap;
select (@after:=unix_timestamp())*0; # always give repeatable output
# verify that innodb_lock_wait_timeout was exceeded. When there was
# the bug, the reap would return immediately after the insert into t2.
select (@after-@before) >= 2;
drop table t1,t2;
commit;
# test for BUG#7947 - DO RELEASE_LOCK() not written to binlog on rollback in the middle
# of a transaction
connection con2;
begin;
create temporary table ti (a int) engine=innodb;
rollback;
insert into ti values(1);
set autocommit=0;
create temporary table t1 (a int) engine=myisam;
commit;
insert t1 values (1);
rollback;
create table t0 (n int);
insert t0 select * from t1;
set autocommit=1;
insert into t0 select GET_LOCK("lock1",null);
set autocommit=0;
create table t2 (n int) engine=innodb;
insert into t2 values (3);
disconnect con2;
connection con3;
select get_lock("lock1",60);
--replace_column 5 #
--replace_result "xid=208" "xid=105" "xid=227" "xid=114" "xid=230" "xid=115" "xid=234" "xid=117" "xid=261" "xid=132"
show binlog events from 98;
do release_lock("lock1");
drop table t0,t2;
Added end marker for tests to make future merges easier mysql-test/t/alias.test: Added end marker for test to make future merges easier mysql-test/t/alter_table.test: Added end marker for test to make future merges easier mysql-test/t/analyse.test: Added end marker for test to make future merges easier mysql-test/t/analyze.test: Added end marker for test to make future merges easier Fixed length of comment lines mysql-test/t/ansi.test: Added end marker for test to make future merges easier mysql-test/t/archive.test: Added end marker for test to make future merges easier mysql-test/t/auto_increment.test: Added end marker for test to make future merges easier mysql-test/t/backup.test: Added end marker for test to make future merges easier mysql-test/t/bdb-alter-table-1.test: Added end marker for test to make future merges easier mysql-test/t/bdb-alter-table-2.test: Added end marker for test to make future merges easier mysql-test/t/bdb-crash.test: Added end marker for test to make future merges easier mysql-test/t/bdb-deadlock.test: Added end marker for test to make future merges easier mysql-test/t/bdb-deadlock.tminus: Added end marker for test to make future merges easier mysql-test/t/bdb.test: Added end marker for test to make future merges easier mysql-test/t/bdb_cache.test: Added end marker for test to make future merges easier mysql-test/t/bench_count_distinct.test: Added end marker for test to make future merges easier mysql-test/t/bigint.test: Added end marker for test to make future merges easier mysql-test/t/binary.test: Added end marker for test to make future merges easier mysql-test/t/blackhole.test: Added end marker for test to make future merges easier mysql-test/t/bool.test: Added end marker for test to make future merges easier mysql-test/t/bulk_replace.test: Added end marker for test to make future merges easier mysql-test/t/case.test: Added end marker for test to make future merges easier mysql-test/t/cast.test: Added end marker for test to make future merges easier mysql-test/t/check.test: Added end marker for test to make future merges easier mysql-test/t/comments.test: Added end marker for test to make future merges easier mysql-test/t/compare.test: Added end marker for test to make future merges easier mysql-test/t/connect.test: Added end marker for test to make future merges easier mysql-test/t/consistent_snapshot.test: Added end marker for test to make future merges easier mysql-test/t/constraints.test: Added end marker for test to make future merges easier mysql-test/t/count_distinct.test: Added end marker for test to make future merges easier mysql-test/t/count_distinct2.test: Added end marker for test to make future merges easier mysql-test/t/count_distinct3.test: Added end marker for test to make future merges easier mysql-test/t/create.test: Added end marker for test to make future merges easier mysql-test/t/create_select_tmp.test: Added end marker for test to make future merges easier mysql-test/t/csv.test: Added end marker for test to make future merges easier mysql-test/t/ctype_big5.test: Added end marker for test to make future merges easier mysql-test/t/ctype_collate.test: Added end marker for test to make future merges easier mysql-test/t/ctype_cp1250_ch.test: Added end marker for test to make future merges easier mysql-test/t/ctype_cp1251.test: Added end marker for test to make future merges easier mysql-test/t/ctype_cp932.test: Added end marker for test to make future merges easier mysql-test/t/ctype_create.test: Added end marker for test to make future merges easier mysql-test/t/ctype_gbk.test: Added end marker for test to make future merges easier mysql-test/t/ctype_latin1.test: Added end marker for test to make future merges easier mysql-test/t/ctype_latin1_de.test: Added end marker for test to make future merges easier mysql-test/t/ctype_latin2.test: Added end marker for test to make future merges easier mysql-test/t/ctype_many.test: Added end marker for test to make future merges easier mysql-test/t/ctype_mb.test: Added end marker for test to make future merges easier mysql-test/t/ctype_recoding.test: Added end marker for test to make future merges easier mysql-test/t/ctype_sjis.test: Added end marker for test to make future merges easier mysql-test/t/ctype_tis620.test: Added end marker for test to make future merges easier mysql-test/t/ctype_uca.test: Added end marker for test to make future merges easier mysql-test/t/ctype_ucs.test: Added end marker for test to make future merges easier mysql-test/t/ctype_ucs_binlog.test: Added end marker for test to make future merges easier mysql-test/t/ctype_ujis.test: Added end marker for test to make future merges easier mysql-test/t/ctype_utf8.test: Added end marker for test to make future merges easier mysql-test/t/date_formats.test: Added end marker for test to make future merges easier mysql-test/t/delayed.test: Added end marker for test to make future merges easier mysql-test/t/delete.test: Added end marker for test to make future merges easier mysql-test/t/derived.test: Added end marker for test to make future merges easier mysql-test/t/dirty_close.test: Added end marker for test to make future merges easier mysql-test/t/distinct.test: Added end marker for test to make future merges easier mysql-test/t/drop.test: Added end marker for test to make future merges easier mysql-test/t/drop_temp_table.test: Added end marker for test to make future merges easier mysql-test/t/empty_table.test: Added end marker for test to make future merges easier mysql-test/t/endspace.test: Added end marker for test to make future merges easier mysql-test/t/errors.test: Added end marker for test to make future merges easier mysql-test/t/exampledb.test: Added end marker for test to make future merges easier mysql-test/t/explain.test: Added end marker for test to make future merges easier mysql-test/t/flush.test: Added end marker for test to make future merges easier mysql-test/t/flush_block_commit.test: Added end marker for test to make future merges easier mysql-test/t/flush_table.test: Added end marker for test to make future merges easier mysql-test/t/foreign_key.test: Added end marker for test to make future merges easier mysql-test/t/fulltext.test: Added end marker for test to make future merges easier mysql-test/t/fulltext2.test: Added end marker for test to make future merges easier mysql-test/t/fulltext_cache.test: Added end marker for test to make future merges easier mysql-test/t/fulltext_distinct.test: Added end marker for test to make future merges easier mysql-test/t/fulltext_left_join.test: Added end marker for test to make future merges easier mysql-test/t/fulltext_multi.test: Added end marker for test to make future merges easier mysql-test/t/fulltext_order_by.test: Added end marker for test to make future merges easier mysql-test/t/fulltext_update.test: Added end marker for test to make future merges easier mysql-test/t/fulltext_var.test: Added end marker for test to make future merges easier mysql-test/t/func_compress.test: Added end marker for test to make future merges easier mysql-test/t/func_concat.test: Added end marker for test to make future merges easier mysql-test/t/func_crypt.test: Added end marker for test to make future merges easier mysql-test/t/func_date_add.test: Added end marker for test to make future merges easier mysql-test/t/func_default.test: Added end marker for test to make future merges easier mysql-test/t/func_des_encrypt.test: Added end marker for test to make future merges easier mysql-test/t/func_encrypt.test: Added end marker for test to make future merges easier mysql-test/t/func_encrypt_nossl.test: Added end marker for test to make future merges easier mysql-test/t/func_equal.test: Added end marker for test to make future merges easier mysql-test/t/func_gconcat.test: Added end marker for test to make future merges easier mysql-test/t/func_group.test: Added end marker for test to make future merges easier mysql-test/t/func_if.test: Added end marker for test to make future merges easier mysql-test/t/func_in.test: Added end marker for test to make future merges easier mysql-test/t/func_isnull.test: Added end marker for test to make future merges easier mysql-test/t/func_like.test: Added end marker for test to make future merges easier mysql-test/t/func_math.test: Added end marker for test to make future merges easier mysql-test/t/func_misc.test: Added end marker for test to make future merges easier mysql-test/t/func_op.test: Added end marker for test to make future merges easier mysql-test/t/func_regexp.test: Added end marker for test to make future merges easier mysql-test/t/func_sapdb.test: Added end marker for test to make future merges easier mysql-test/t/func_set.test: Added end marker for test to make future merges easier mysql-test/t/func_str.test: Added end marker for test to make future merges easier mysql-test/t/func_system.test: Added end marker for test to make future merges easier mysql-test/t/func_test.test: Added end marker for test to make future merges easier mysql-test/t/func_time.test: Added end marker for test to make future merges easier mysql-test/t/func_timestamp.test: Added end marker for test to make future merges easier mysql-test/t/gcc296.test: Added end marker for test to make future merges easier mysql-test/t/gis-rtree.test: Added end marker for test to make future merges easier mysql-test/t/gis.test: Added end marker for test to make future merges easier mysql-test/t/grant.test: Added end marker for test to make future merges easier mysql-test/t/grant2.test: Added end marker for test to make future merges easier mysql-test/t/grant_cache.test: Added end marker for test to make future merges easier mysql-test/t/group_by.test: Added end marker for test to make future merges easier mysql-test/t/handler.test: Added end marker for test to make future merges easier mysql-test/t/having.test: Added end marker for test to make future merges easier mysql-test/t/heap.test: Added end marker for test to make future merges easier mysql-test/t/heap_auto_increment.test: Added end marker for test to make future merges easier mysql-test/t/heap_btree.test: Added end marker for test to make future merges easier mysql-test/t/heap_hash.test: Added end marker for test to make future merges easier mysql-test/t/help.test: Added end marker for test to make future merges easier mysql-test/t/init_connect.test: Added end marker for test to make future merges easier mysql-test/t/init_file.test: Added end marker for test to make future merges easier mysql-test/t/innodb-deadlock.test: Added end marker for test to make future merges easier mysql-test/t/innodb-lock.test: Added end marker for test to make future merges easier mysql-test/t/innodb-replace.test: Added end marker for test to make future merges easier mysql-test/t/innodb.test: Added end marker for test to make future merges easier mysql-test/t/innodb_cache.test: Added end marker for test to make future merges easier mysql-test/t/innodb_handler.test: Added end marker for test to make future merges easier mysql-test/t/insert.test: Added end marker for test to make future merges easier mysql-test/t/insert_select-binlog.test: Added end marker for test to make future merges easier mysql-test/t/insert_select.test: Added end marker for test to make future merges easier mysql-test/t/insert_update.test: Added end marker for test to make future merges easier mysql-test/t/isam.test: Added end marker for test to make future merges easier mysql-test/t/join.test: Added end marker for test to make future merges easier mysql-test/t/join_crash.test: Added end marker for test to make future merges easier mysql-test/t/join_outer.test: Added end marker for test to make future merges easier mysql-test/t/key.test: Added end marker for test to make future merges easier mysql-test/t/key_cache.test: Added end marker for test to make future merges easier mysql-test/t/key_diff.test: Added end marker for test to make future merges easier mysql-test/t/key_primary.test: Added end marker for test to make future merges easier mysql-test/t/keywords.test: Added end marker for test to make future merges easier mysql-test/t/kill.test: Added end marker for test to make future merges easier mysql-test/t/limit.test: Added end marker for test to make future merges easier mysql-test/t/loaddata.test: Added end marker for test to make future merges easier mysql-test/t/lock.test: Added end marker for test to make future merges easier mysql-test/t/lock_multi.test: Added end marker for test to make future merges easier mysql-test/t/lock_tables_lost_commit.test: Added end marker for test to make future merges easier mysql-test/t/lowercase_table.test: Added end marker for test to make future merges easier mysql-test/t/lowercase_table2.test: Added end marker for test to make future merges easier mysql-test/t/lowercase_table3.test: Added end marker for test to make future merges easier mysql-test/t/lowercase_table_grant.test: Added end marker for test to make future merges easier mysql-test/t/lowercase_table_qcache.test: Added end marker for test to make future merges easier mysql-test/t/merge.test: Added end marker for test to make future merges easier mysql-test/t/metadata.test: Added end marker for test to make future merges easier mysql-test/t/mix_innodb_myisam_binlog.test: Added end marker for test to make future merges easier mysql-test/t/multi_statement.test: Added end marker for test to make future merges easier mysql-test/t/multi_update.test: Added end marker for test to make future merges easier mysql-test/t/myisam-blob.test: Added end marker for test to make future merges easier mysql-test/t/myisam.test: Added end marker for test to make future merges easier mysql-test/t/mysql_client_test.test: Added end marker for test to make future merges easier mysql-test/t/mysql_protocols.test: Added end marker for test to make future merges easier mysql-test/t/mysqlbinlog.test: Added end marker for test to make future merges easier mysql-test/t/mysqlbinlog2.test: Added end marker for test to make future merges easier mysql-test/t/mysqldump.test: Added end marker for test to make future merges easier mysql-test/t/mysqltest.test: Added end marker for test to make future merges easier mysql-test/t/ndb_alter_table.test: Added end marker for test to make future merges easier mysql-test/t/ndb_autodiscover.test: Added end marker for test to make future merges easier mysql-test/t/ndb_autodiscover2.test: Added end marker for test to make future merges easier mysql-test/t/ndb_basic.test: Added end marker for test to make future merges easier mysql-test/t/ndb_blob.test: Added end marker for test to make future merges easier mysql-test/t/ndb_cache.test: Added end marker for test to make future merges easier mysql-test/t/ndb_charset.test: Added end marker for test to make future merges easier mysql-test/t/ndb_config.test: Added end marker for test to make future merges easier mysql-test/t/ndb_database.test: Added end marker for test to make future merges easier mysql-test/t/ndb_grant.later: Added end marker for test to make future merges easier mysql-test/t/ndb_index.test: Added end marker for test to make future merges easier mysql-test/t/ndb_index_ordered.test: Added end marker for test to make future merges easier mysql-test/t/ndb_index_unique.test: Added end marker for test to make future merges easier mysql-test/t/ndb_insert.test: Added end marker for test to make future merges easier mysql-test/t/ndb_limit.test: Added end marker for test to make future merges easier mysql-test/t/ndb_lock.test: Added end marker for test to make future merges easier mysql-test/t/ndb_minmax.test: Added end marker for test to make future merges easier mysql-test/t/ndb_multi.test: Added end marker for test to make future merges easier mysql-test/t/ndb_replace.test: Added end marker for test to make future merges easier mysql-test/t/ndb_restore.test: Added end marker for test to make future merges easier mysql-test/t/ndb_subquery.test: Added end marker for test to make future merges easier mysql-test/t/ndb_transaction.test: Added end marker for test to make future merges easier mysql-test/t/ndb_truncate.test: Added end marker for test to make future merges easier mysql-test/t/ndb_types.test: Added end marker for test to make future merges easier mysql-test/t/ndb_update.test: Added end marker for test to make future merges easier mysql-test/t/negation_elimination.test: Added end marker for test to make future merges easier mysql-test/t/not_embedded_server.test: Added end marker for test to make future merges easier mysql-test/t/null.test: Added end marker for test to make future merges easier mysql-test/t/null_key.test: Added end marker for test to make future merges easier mysql-test/t/odbc.test: Added end marker for test to make future merges easier mysql-test/t/olap.test: Added end marker for test to make future merges easier mysql-test/t/openssl_1.test: Added end marker for test to make future merges easier mysql-test/t/order_by.test: Added end marker for test to make future merges easier mysql-test/t/order_fill_sortbuf.test: Added end marker for test to make future merges easier mysql-test/t/outfile.test: Added end marker for test to make future merges easier mysql-test/t/overflow.test: Added end marker for test to make future merges easier mysql-test/t/packet.test: Added end marker for test to make future merges easier mysql-test/t/preload.test: Added end marker for test to make future merges easier mysql-test/t/ps.test: Added end marker for test to make future merges easier mysql-test/t/ps_10nestset.test: Added end marker for test to make future merges easier mysql-test/t/ps_11bugs.test: Added end marker for test to make future merges easier mysql-test/t/ps_1general.test: Added end marker for test to make future merges easier mysql-test/t/ps_2myisam.test: Added end marker for test to make future merges easier mysql-test/t/ps_3innodb.test: Added end marker for test to make future merges easier mysql-test/t/ps_4heap.test: Added end marker for test to make future merges easier mysql-test/t/ps_5merge.test: Added end marker for test to make future merges easier mysql-test/t/ps_6bdb.test: Added end marker for test to make future merges easier mysql-test/t/ps_7ndb.test: Added end marker for test to make future merges easier mysql-test/t/ps_grant.test: Added end marker for test to make future merges easier mysql-test/t/query_cache.test: Added end marker for test to make future merges easier mysql-test/t/query_cache_merge.test: Added end marker for test to make future merges easier mysql-test/t/raid.test: Added end marker for test to make future merges easier mysql-test/t/range.test: Added end marker for test to make future merges easier mysql-test/t/rename.test: Added end marker for test to make future merges easier mysql-test/t/repair.test: Added end marker for test to make future merges easier mysql-test/t/replace.test: Added end marker for test to make future merges easier mysql-test/t/rollback.test: Added end marker for test to make future merges easier mysql-test/t/row.test: Added end marker for test to make future merges easier mysql-test/t/rpl000001.test: Added end marker for test to make future merges easier mysql-test/t/rpl000002.test: Added end marker for test to make future merges easier mysql-test/t/rpl000004.test: Added end marker for test to make future merges easier mysql-test/t/rpl000005.test: Added end marker for test to make future merges easier mysql-test/t/rpl000006.test: Added end marker for test to make future merges easier mysql-test/t/rpl000008.test: Added end marker for test to make future merges easier mysql-test/t/rpl000009.test: Added end marker for test to make future merges easier mysql-test/t/rpl000010.test: Added end marker for test to make future merges easier mysql-test/t/rpl000011.test: Added end marker for test to make future merges easier mysql-test/t/rpl000012.test: Added end marker for test to make future merges easier mysql-test/t/rpl000013.test: Added end marker for test to make future merges easier mysql-test/t/rpl000015.test: Added end marker for test to make future merges easier mysql-test/t/rpl000017.test: Added end marker for test to make future merges easier mysql-test/t/rpl000018.test: Added end marker for test to make future merges easier mysql-test/t/rpl_EE_error.test: Added end marker for test to make future merges easier mysql-test/t/rpl_alter.test: Added end marker for test to make future merges easier mysql-test/t/rpl_chain_temp_table.test: Added end marker for test to make future merges easier mysql-test/t/rpl_change_master.test: Added end marker for test to make future merges easier mysql-test/t/rpl_charset.test: Added end marker for test to make future merges easier mysql-test/t/rpl_commit_after_flush.test: Added end marker for test to make future merges easier mysql-test/t/rpl_create_database.test: Added end marker for test to make future merges easier mysql-test/t/rpl_ddl.test: Added end marker for test to make future merges easier mysql-test/t/rpl_deadlock.test: Added end marker for test to make future merges easier mysql-test/t/rpl_delete_all.test: Added end marker for test to make future merges easier mysql-test/t/rpl_do_grant.test: Added end marker for test to make future merges easier mysql-test/t/rpl_drop.test: Added end marker for test to make future merges easier mysql-test/t/rpl_drop_temp.test: Added end marker for test to make future merges easier mysql-test/t/rpl_empty_master_crash.test: Added end marker for test to make future merges easier mysql-test/t/rpl_error_ignored_table.test: Added end marker for test to make future merges easier mysql-test/t/rpl_failed_optimize.test: Added end marker for test to make future merges easier mysql-test/t/rpl_failsafe.test: Added end marker for test to make future merges easier mysql-test/t/rpl_flush_log_loop.test: Added end marker for test to make future merges easier mysql-test/t/rpl_flush_tables.test: Added end marker for test to make future merges easier mysql-test/t/rpl_free_items.test: Added end marker for test to make future merges easier mysql-test/t/rpl_get_lock.test: Added end marker for test to make future merges easier mysql-test/t/rpl_heap.test: Added end marker for test to make future merges easier mysql-test/t/rpl_ignore_grant.test: Added end marker for test to make future merges easier mysql-test/t/rpl_init_slave.test: Added end marker for test to make future merges easier mysql-test/t/rpl_innodb.test: Added end marker for test to make future merges easier mysql-test/t/rpl_insert_id.test: Added end marker for test to make future merges easier mysql-test/t/rpl_insert_ignore.test: Added end marker for test to make future merges easier mysql-test/t/rpl_loaddata.test: Added end marker for test to make future merges easier mysql-test/t/rpl_loaddata_rule_m.test: Added end marker for test to make future merges easier mysql-test/t/rpl_loaddata_rule_s.test: Added end marker for test to make future merges easier mysql-test/t/rpl_loaddatalocal.test: Added end marker for test to make future merges easier mysql-test/t/rpl_log.test: Added end marker for test to make future merges easier mysql-test/t/rpl_log_pos.test: Added end marker for test to make future merges easier mysql-test/t/rpl_many_optimize.test: Added end marker for test to make future merges easier mysql-test/t/rpl_master_pos_wait.test: Added end marker for test to make future merges easier mysql-test/t/rpl_max_relay_size.test: Added end marker for test to make future merges easier mysql-test/t/rpl_misc_functions.test: Added end marker for test to make future merges easier mysql-test/t/rpl_multi_delete.test: Added end marker for test to make future merges easier mysql-test/t/rpl_multi_delete2.test: Added end marker for test to make future merges easier mysql-test/t/rpl_multi_query.test: Added end marker for test to make future merges easier mysql-test/t/rpl_multi_update.test: Added end marker for test to make future merges easier mysql-test/t/rpl_multi_update2.test: Added end marker for test to make future merges easier mysql-test/t/rpl_multi_update3.test: Added end marker for test to make future merges easier mysql-test/t/rpl_mystery22.test: Added end marker for test to make future merges easier mysql-test/t/rpl_openssl.test: Added end marker for test to make future merges easier mysql-test/t/rpl_optimize.test: Added end marker for test to make future merges easier mysql-test/t/rpl_ps.test: Added end marker for test to make future merges easier mysql-test/t/rpl_redirect.test: Added end marker for test to make future merges easier mysql-test/t/rpl_relayrotate.test: Added end marker for test to make future merges easier mysql-test/t/rpl_relayspace.test: Added end marker for test to make future merges easier mysql-test/t/rpl_replicate_do.test: Added end marker for test to make future merges easier mysql-test/t/rpl_reset_slave.test: Added end marker for test to make future merges easier mysql-test/t/rpl_rewrite_db.test: Added end marker for test to make future merges easier mysql-test/t/rpl_rotate_logs.test: Added end marker for test to make future merges easier mysql-test/t/rpl_server_id1.test: Added end marker for test to make future merges easier mysql-test/t/rpl_server_id2.test: Added end marker for test to make future merges easier mysql-test/t/rpl_set_charset.test: Added end marker for test to make future merges easier mysql-test/t/rpl_skip_error.test: Added end marker for test to make future merges easier mysql-test/t/rpl_sporadic_master.test: Added end marker for test to make future merges easier mysql-test/t/rpl_start_stop_slave.test: Added end marker for test to make future merges easier mysql-test/t/rpl_temporary.test: Added end marker for test to make future merges easier mysql-test/t/rpl_timezone.test: Added end marker for test to make future merges easier mysql-test/t/rpl_trunc_binlog.test: Added end marker for test to make future merges easier mysql-test/t/rpl_until.test: Added end marker for test to make future merges easier mysql-test/t/rpl_user_variables.test: Added end marker for test to make future merges easier mysql-test/t/rpl_variables.test: Added end marker for test to make future merges easier mysql-test/t/select.test: Added end marker for test to make future merges easier mysql-test/t/select_found.test: Added end marker for test to make future merges easier mysql-test/t/select_safe.test: Added end marker for test to make future merges easier mysql-test/t/show_check.test: Added end marker for test to make future merges easier mysql-test/t/skip_name_resolve.test: Added end marker for test to make future merges easier mysql-test/t/sql_mode.test: Added end marker for test to make future merges easier mysql-test/t/status.test: Added end marker for test to make future merges easier mysql-test/t/subselect.test: Added end marker for test to make future merges easier mysql-test/t/subselect2.test: Added end marker for test to make future merges easier mysql-test/t/subselect_gis.test: Added end marker for test to make future merges easier mysql-test/t/subselect_innodb.test: Added end marker for test to make future merges easier mysql-test/t/symlink.test: Added end marker for test to make future merges easier mysql-test/t/synchronization.test: Added end marker for test to make future merges easier mysql-test/t/system_mysql_db.test: Added end marker for test to make future merges easier mysql-test/t/system_mysql_db_fix.test: Added end marker for test to make future merges easier mysql-test/t/system_mysql_db_refs.test: Added end marker for test to make future merges easier mysql-test/t/tablelock.test: Added end marker for test to make future merges easier mysql-test/t/temp_table.test: Added end marker for test to make future merges easier mysql-test/t/timezone.test: Added end marker for test to make future merges easier mysql-test/t/timezone2.test: Added end marker for test to make future merges easier mysql-test/t/timezone3.test: Added end marker for test to make future merges easier mysql-test/t/timezone_grant.test: Added end marker for test to make future merges easier mysql-test/t/truncate.test: Added end marker for test to make future merges easier mysql-test/t/type_blob.test: Added end marker for test to make future merges easier mysql-test/t/type_date.test: Added end marker for test to make future merges easier mysql-test/t/type_datetime.test: Added end marker for test to make future merges easier mysql-test/t/type_decimal.test: Added end marker for test to make future merges easier mysql-test/t/type_enum.test: Added end marker for test to make future merges easier mysql-test/t/type_float.test: Added end marker for test to make future merges easier mysql-test/t/type_nchar.test: Added end marker for test to make future merges easier mysql-test/t/type_ranges.test: Added end marker for test to make future merges easier mysql-test/t/type_set.test: Added end marker for test to make future merges easier mysql-test/t/type_time.test: Added end marker for test to make future merges easier mysql-test/t/type_timestamp.test: Added end marker for test to make future merges easier mysql-test/t/type_uint.test: Added end marker for test to make future merges easier mysql-test/t/type_year.test: Added end marker for test to make future merges easier mysql-test/t/union.test: Added end marker for test to make future merges easier mysql-test/t/update.test: Added end marker for test to make future merges easier mysql-test/t/user_var-binlog.test: Added end marker for test to make future merges easier mysql-test/t/user_var.test: Added end marker for test to make future merges easier mysql-test/t/varbinary.test: Added end marker for test to make future merges easier mysql-test/t/variables.test: Added end marker for test to make future merges easier mysql-test/t/warnings.test: Added end marker for test to make future merges easier
2005-07-28 02:22:47 +02:00
# End of 4.1 tests
# Test for BUG#16559 (ROLLBACK should always have a zero error code in
# binlog). Has to be here and not earlier, as the SELECTs influence
# XIDs differently between normal and ps-protocol (and SHOW BINLOG
# EVENTS above read XIDs).
connect (con4,localhost,root,,);
connection con3;
reset master;
create table t1 (a int) engine=innodb;
create table t2 (a int) engine=myisam;
select get_lock("a",10);
begin;
insert into t1 values(8);
insert into t2 select * from t1;
disconnect con3;
connection con4;
select get_lock("a",10); # wait for rollback to finish
# we check that the error code of the "ROLLBACK" event is 0 and not
# ER_SERVER_SHUTDOWN (i.e. disconnection just rolls back transaction
# and does not make slave to stop)
--exec $MYSQL_BINLOG --start-position=547 $MYSQLTEST_VARDIR/log/master-bin.000001 > $MYSQLTEST_VARDIR/tmp/mix_innodb_myisam_binlog.output
--replace_result $MYSQLTEST_VARDIR MYSQLTEST_VARDIR
eval select
(@a:=load_file("$MYSQLTEST_VARDIR/tmp/mix_innodb_myisam_binlog.output"))
is not null;
--replace_result $MYSQL_TEST_DIR MYSQL_TEST_DIR
eval select
@a like "%#%error_code=0%ROLLBACK;%ROLLBACK /* added by mysqlbinlog */;%",
@a not like "%#%error_code=%error_code=%";
drop table t1, t2;