2012-03-28 01:04:46 +02:00
|
|
|
/* Copyright (c) 2006, 2012, Oracle and/or its affiliates.
|
fixes for test failures
and small collateral changes
mysql-test/lib/My/Test.pm:
somehow with "print" we get truncated writes sometimes
mysql-test/suite/perfschema/r/digest_table_full.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/dml_handler.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/information_schema.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/nesting.result:
this differs, because we don't rewrite general log queries, and multi-statement
packets are logged as a one entry. this result file is identical to what mysql-5.6.5
produces with the --log-raw option.
mysql-test/suite/perfschema/r/relaylog.result:
MariaDB modifies the binlog index file directly, while MySQL 5.6 has a feature "crash-safe binlog index" and modifies a special "crash-safe" shadow copy of the index file and then moves it over. That's why this test shows "NONE" index file writes in MySQL and "MANY" in MariaDB.
mysql-test/suite/perfschema/r/server_init.result:
MariaDB initializes the "manager" resources from the "manager" thread, and starts this thread only when --flush-time is not 0. MySQL 5.6 initializes "manager" resources unconditionally on server startup.
mysql-test/suite/perfschema/r/stage_mdl_global.result:
this differs, because MariaDB disables query cache when query_cache_size=0. MySQL does not
do that, and this causes useless mutex locks and waits.
mysql-test/suite/perfschema/r/statement_digest.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_consumers.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_long_query.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/rpl/r/rpl_mixed_drop_create_temp_table.result:
will be updated to match 5.6 when alfranio.correia@oracle.com-20110512172919-c1b5kmum4h52g0ni and anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y are merged
mysql-test/suite/rpl/r/rpl_non_direct_mixed_mixing_engines.result:
will be updated to match 5.6 when anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y is merged
2012-09-27 20:09:46 +02:00
|
|
|
Copyright (c) 2010, 2012, Monty Program Ab.
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
This program is free software; you can redistribute it and/or modify
|
|
|
|
it under the terms of the GNU General Public License as published by
|
2006-12-27 02:23:51 +01:00
|
|
|
the Free Software Foundation; version 2 of the License.
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
This program is distributed in the hope that it will be useful,
|
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
GNU General Public License for more details.
|
|
|
|
|
|
|
|
You should have received a copy of the GNU General Public License
|
fixes for test failures
and small collateral changes
mysql-test/lib/My/Test.pm:
somehow with "print" we get truncated writes sometimes
mysql-test/suite/perfschema/r/digest_table_full.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/dml_handler.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/information_schema.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/nesting.result:
this differs, because we don't rewrite general log queries, and multi-statement
packets are logged as a one entry. this result file is identical to what mysql-5.6.5
produces with the --log-raw option.
mysql-test/suite/perfschema/r/relaylog.result:
MariaDB modifies the binlog index file directly, while MySQL 5.6 has a feature "crash-safe binlog index" and modifies a special "crash-safe" shadow copy of the index file and then moves it over. That's why this test shows "NONE" index file writes in MySQL and "MANY" in MariaDB.
mysql-test/suite/perfschema/r/server_init.result:
MariaDB initializes the "manager" resources from the "manager" thread, and starts this thread only when --flush-time is not 0. MySQL 5.6 initializes "manager" resources unconditionally on server startup.
mysql-test/suite/perfschema/r/stage_mdl_global.result:
this differs, because MariaDB disables query cache when query_cache_size=0. MySQL does not
do that, and this causes useless mutex locks and waits.
mysql-test/suite/perfschema/r/statement_digest.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_consumers.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_long_query.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/rpl/r/rpl_mixed_drop_create_temp_table.result:
will be updated to match 5.6 when alfranio.correia@oracle.com-20110512172919-c1b5kmum4h52g0ni and anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y are merged
mysql-test/suite/rpl/r/rpl_non_direct_mixed_mixing_engines.result:
will be updated to match 5.6 when anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y is merged
2012-09-27 20:09:46 +02:00
|
|
|
along with this program; if not, write to the Free Software Foundation,
|
|
|
|
51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA */
|
2006-10-31 12:23:14 +01:00
|
|
|
|
2010-03-31 16:05:33 +02:00
|
|
|
#include "sql_priv.h"
|
|
|
|
#include "unireg.h" // HAVE_*
|
2007-04-12 08:58:04 +02:00
|
|
|
#include "rpl_mi.h"
|
2006-10-31 12:23:14 +01:00
|
|
|
#include "rpl_rli.h"
|
2010-03-31 16:05:33 +02:00
|
|
|
#include "sql_base.h" // close_thread_tables
|
2006-10-31 12:23:14 +01:00
|
|
|
#include <my_dir.h> // For MY_STAT
|
|
|
|
#include "sql_repl.h" // For check_binlog_magic
|
2010-03-31 16:05:33 +02:00
|
|
|
#include "log_event.h" // Format_description_log_event, Log_event,
|
|
|
|
// FORMAT_DESCRIPTION_LOG_EVENT, ROTATE_EVENT,
|
|
|
|
// PREFIX_SQL_LOAD
|
2007-02-26 17:44:55 +01:00
|
|
|
#include "rpl_utility.h"
|
2009-12-03 19:37:38 +01:00
|
|
|
#include "transaction.h"
|
2010-03-31 16:05:33 +02:00
|
|
|
#include "sql_parse.h" // end_trans, ROLLBACK
|
2010-10-28 16:31:21 +02:00
|
|
|
#include <mysql/plugin.h>
|
|
|
|
#include <mysql/service_thd_wait.h>
|
2006-10-31 12:23:14 +01:00
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
static int count_relay_log_space(Relay_log_info* rli);
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
// Defined in slave.cc
|
|
|
|
int init_intvar_from_file(int* var, IO_CACHE* f, int default_val);
|
|
|
|
int init_strvar_from_file(char *var, int max_size, IO_CACHE *f,
|
|
|
|
const char *default_val);
|
|
|
|
|
BUG#40337 Fsyncing master and relay log to disk after every event is too slow
NOTE: Backporting the patch to next-mr.
The fix proposed in BUG#35542 and BUG#31665 introduces a performance issue
when fsyncing the master.info, relay.info and relay-log.bin* after #th events.
Although such solution has been proposed to reduce the probability of corrupted
files due to a slave-crash, the performance penalty introduced by it has
made the approach impractical for highly intensive workloads.
In a nutshell, the option --syn-relay-log proposed in BUG#35542 and BUG#31665
simultaneously fsyncs master.info, relay-log.info and relay-log.bin* and
this is the main source of performance issues.
This patch introduces new options that give more control to the user on
what should be fsynced and how often:
1) (--sync-master-info, integer) which syncs the master.info after #th event;
2) (--sync-relay-log, integer) which syncs the relay-log.bin* after #th
events.
3) (--sync-relay-log-info, integer) which syncs the relay.info after #th
transactions.
To provide both performance and increased reliability, we recommend the following
setup:
1) --sync-master-info = 0 eventually the operating system will fsync it;
2) --sync-relay-log = 0 eventually the operating system will fsync it;
3) --sync-relay-log-info = 1 fsyncs it after every transaction;
Notice, that the previous setup does not reduce the probability of
corrupted master.info and relay-log.bin*. To overcome the issue, this patch also
introduces a recovery mechanism that right after restart throws away relay-log.bin*
retrieved from a master and updates the master.info based on the relay.info:
4) (--relay-log-recovery, boolean) which enables a recovery mechanism that
throws away relay-log.bin* after a crash.
However, it can only recover the incorrect binlog file and position in master.info,
if other informations (host, port password, etc) are corrupted or incorrect,
then this recovery mechanism will fail to work.
2009-09-29 16:40:52 +02:00
|
|
|
Relay_log_info::Relay_log_info(bool is_slave_recovery)
|
2007-06-09 07:19:37 +02:00
|
|
|
:Slave_reporting_capability("SQL"),
|
2007-06-09 08:29:51 +02:00
|
|
|
no_storage(FALSE), replicate_same_server_id(::replicate_same_server_id),
|
2009-09-29 16:27:12 +02:00
|
|
|
info_fd(-1), cur_log_fd(-1), relay_log(&sync_relaylog_period),
|
BUG#40337 Fsyncing master and relay log to disk after every event is too slow
NOTE: Backporting the patch to next-mr.
The fix proposed in BUG#35542 and BUG#31665 introduces a performance issue
when fsyncing the master.info, relay.info and relay-log.bin* after #th events.
Although such solution has been proposed to reduce the probability of corrupted
files due to a slave-crash, the performance penalty introduced by it has
made the approach impractical for highly intensive workloads.
In a nutshell, the option --syn-relay-log proposed in BUG#35542 and BUG#31665
simultaneously fsyncs master.info, relay-log.info and relay-log.bin* and
this is the main source of performance issues.
This patch introduces new options that give more control to the user on
what should be fsynced and how often:
1) (--sync-master-info, integer) which syncs the master.info after #th event;
2) (--sync-relay-log, integer) which syncs the relay-log.bin* after #th
events.
3) (--sync-relay-log-info, integer) which syncs the relay.info after #th
transactions.
To provide both performance and increased reliability, we recommend the following
setup:
1) --sync-master-info = 0 eventually the operating system will fsync it;
2) --sync-relay-log = 0 eventually the operating system will fsync it;
3) --sync-relay-log-info = 1 fsyncs it after every transaction;
Notice, that the previous setup does not reduce the probability of
corrupted master.info and relay-log.bin*. To overcome the issue, this patch also
introduces a recovery mechanism that right after restart throws away relay-log.bin*
retrieved from a master and updates the master.info based on the relay.info:
4) (--relay-log-recovery, boolean) which enables a recovery mechanism that
throws away relay-log.bin* after a crash.
However, it can only recover the incorrect binlog file and position in master.info,
if other informations (host, port password, etc) are corrupted or incorrect,
then this recovery mechanism will fail to work.
2009-09-29 16:40:52 +02:00
|
|
|
sync_counter(0), is_relay_log_recovery(is_slave_recovery),
|
2009-10-23 00:30:28 +02:00
|
|
|
save_temporary_tables(0), cur_log_old_open_count(0), group_relay_log_pos(0),
|
|
|
|
event_relay_log_pos(0),
|
2009-05-06 14:03:24 +02:00
|
|
|
#if HAVE_valgrind
|
2008-02-11 12:04:30 +01:00
|
|
|
is_fake(FALSE),
|
|
|
|
#endif
|
2008-02-18 12:35:44 +01:00
|
|
|
group_master_log_pos(0), log_space_total(0), ignore_log_space_limit(0),
|
|
|
|
last_master_timestamp(0), slave_skip_counter(0),
|
2007-06-09 07:19:37 +02:00
|
|
|
abort_pos_wait(0), slave_run_id(0), sql_thd(0),
|
2006-10-31 12:23:14 +01:00
|
|
|
inited(0), abort_slave(0), slave_running(0), until_condition(UNTIL_NONE),
|
2012-10-03 00:44:54 +02:00
|
|
|
until_log_pos(0), retried_trans(0), executed_entries(0),
|
2006-10-31 12:23:14 +01:00
|
|
|
tables_to_lock(0), tables_to_lock_count(0),
|
2012-04-21 12:24:39 +02:00
|
|
|
last_event_start_time(0), deferred_events(NULL),m_flags(0),
|
2012-06-14 20:05:31 +02:00
|
|
|
row_stmt_start_timestamp(0), long_find_row_note_printed(false),
|
2011-01-10 14:53:09 +01:00
|
|
|
m_annotate_event(0)
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::Relay_log_info");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
2012-12-14 15:38:07 +01:00
|
|
|
relay_log.is_relay_log= TRUE;
|
2011-03-01 17:39:28 +01:00
|
|
|
#ifdef HAVE_PSI_INTERFACE
|
|
|
|
relay_log.set_psi_keys(key_RELAYLOG_LOCK_index,
|
|
|
|
key_RELAYLOG_update_cond,
|
|
|
|
key_file_relaylog,
|
2011-10-25 12:53:40 +02:00
|
|
|
key_file_relaylog_index,
|
|
|
|
key_RELAYLOG_COND_queue_busy);
|
2011-03-01 17:39:28 +01:00
|
|
|
#endif
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
group_relay_log_name[0]= event_relay_log_name[0]=
|
|
|
|
group_master_log_name[0]= 0;
|
2007-06-09 07:19:37 +02:00
|
|
|
until_log_name[0]= ign_master_log_name_end[0]= 0;
|
2012-10-01 01:30:44 +02:00
|
|
|
max_relay_log_size= global_system_variables.max_relay_log_size;
|
2006-10-31 12:23:14 +01:00
|
|
|
bzero((char*) &info_file, sizeof(info_file));
|
|
|
|
bzero((char*) &cache_buf, sizeof(cache_buf));
|
|
|
|
cached_charset_invalidate();
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_init(key_relay_log_info_run_lock, &run_lock, MY_MUTEX_INIT_FAST);
|
|
|
|
mysql_mutex_init(key_relay_log_info_data_lock,
|
|
|
|
&data_lock, MY_MUTEX_INIT_FAST);
|
|
|
|
mysql_mutex_init(key_relay_log_info_log_space_lock,
|
|
|
|
&log_space_lock, MY_MUTEX_INIT_FAST);
|
2012-01-23 13:09:37 +01:00
|
|
|
mysql_mutex_init(key_relay_log_info_sleep_lock, &sleep_lock, MY_MUTEX_INIT_FAST);
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_cond_init(key_relay_log_info_data_cond, &data_cond, NULL);
|
|
|
|
mysql_cond_init(key_relay_log_info_start_cond, &start_cond, NULL);
|
|
|
|
mysql_cond_init(key_relay_log_info_stop_cond, &stop_cond, NULL);
|
|
|
|
mysql_cond_init(key_relay_log_info_log_space_cond, &log_space_cond, NULL);
|
2012-01-23 13:09:37 +01:00
|
|
|
mysql_cond_init(key_relay_log_info_sleep_cond, &sleep_cond, NULL);
|
2006-10-31 12:23:14 +01:00
|
|
|
relay_log.init_pthread_objects();
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
Relay_log_info::~Relay_log_info()
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::~Relay_log_info");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_destroy(&run_lock);
|
|
|
|
mysql_mutex_destroy(&data_lock);
|
|
|
|
mysql_mutex_destroy(&log_space_lock);
|
2012-01-23 13:09:37 +01:00
|
|
|
mysql_mutex_destroy(&sleep_lock);
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_cond_destroy(&data_cond);
|
|
|
|
mysql_cond_destroy(&start_cond);
|
|
|
|
mysql_cond_destroy(&stop_cond);
|
|
|
|
mysql_cond_destroy(&log_space_cond);
|
2012-01-23 13:09:37 +01:00
|
|
|
mysql_cond_destroy(&sleep_cond);
|
2006-10-31 12:23:14 +01:00
|
|
|
relay_log.cleanup();
|
2011-01-10 14:53:09 +01:00
|
|
|
free_annotate_event();
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
int init_relay_log_info(Relay_log_info* rli,
|
2006-10-31 12:23:14 +01:00
|
|
|
const char* info_fname)
|
|
|
|
{
|
|
|
|
char fname[FN_REFLEN+128];
|
|
|
|
int info_fd;
|
|
|
|
const char* msg = 0;
|
|
|
|
int error = 0;
|
|
|
|
DBUG_ENTER("init_relay_log_info");
|
|
|
|
DBUG_ASSERT(!rli->no_storage); // Don't init if there is no storage
|
|
|
|
|
|
|
|
if (rli->inited) // Set if this function called
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
fn_format(fname, info_fname, mysql_data_home, "", 4+32);
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_lock(&rli->data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
info_fd = rli->info_fd;
|
|
|
|
rli->cur_log_fd = -1;
|
|
|
|
rli->slave_skip_counter=0;
|
|
|
|
rli->abort_pos_wait=0;
|
|
|
|
rli->log_space_limit= relay_log_space_limit;
|
|
|
|
rli->log_space_total= 0;
|
|
|
|
rli->tables_to_lock= 0;
|
|
|
|
rli->tables_to_lock_count= 0;
|
|
|
|
|
2009-04-19 03:21:33 +02:00
|
|
|
char pattern[FN_REFLEN];
|
2009-11-28 05:43:16 +01:00
|
|
|
(void) my_realpath(pattern, slave_load_tmpdir, 0);
|
|
|
|
if (fn_format(pattern, PREFIX_SQL_LOAD, pattern, "",
|
2009-04-19 03:21:33 +02:00
|
|
|
MY_SAFE_PATH | MY_RETURN_REAL_PATH) == NullS)
|
|
|
|
{
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2009-04-19 03:21:33 +02:00
|
|
|
sql_print_error("Unable to use slave's temporary directory %s",
|
|
|
|
slave_load_tmpdir);
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
unpack_filename(rli->slave_patternload_file, pattern);
|
2009-02-21 10:36:07 +01:00
|
|
|
rli->slave_patternload_file_size= strlen(rli->slave_patternload_file);
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
/*
|
|
|
|
The relay log will now be opened, as a SEQ_READ_APPEND IO_CACHE.
|
|
|
|
Note that the I/O thread flushes it to disk after writing every
|
2010-02-03 17:56:17 +01:00
|
|
|
event, in flush_master_info(mi, 1, ?).
|
2006-10-31 12:23:14 +01:00
|
|
|
*/
|
|
|
|
|
|
|
|
{
|
2009-11-05 07:07:31 +01:00
|
|
|
/* Reports an error and returns, if the --relay-log's path
|
|
|
|
is a directory.*/
|
|
|
|
if (opt_relay_logname &&
|
|
|
|
opt_relay_logname[strlen(opt_relay_logname) - 1] == FN_LIBCHAR)
|
|
|
|
{
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2009-11-05 07:07:31 +01:00
|
|
|
sql_print_error("Path '%s' is a directory name, please specify \
|
|
|
|
a file name for --relay-log option", opt_relay_logname);
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Reports an error and returns, if the --relay-log-index's path
|
|
|
|
is a directory.*/
|
|
|
|
if (opt_relaylog_index_name &&
|
|
|
|
opt_relaylog_index_name[strlen(opt_relaylog_index_name) - 1]
|
|
|
|
== FN_LIBCHAR)
|
|
|
|
{
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2009-11-05 07:07:31 +01:00
|
|
|
sql_print_error("Path '%s' is a directory name, please specify \
|
|
|
|
a file name for --relay-log-index option", opt_relaylog_index_name);
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
char buf[FN_REFLEN];
|
|
|
|
const char *ln;
|
|
|
|
static bool name_warning_sent= 0;
|
|
|
|
ln= rli->relay_log.generate_name(opt_relay_logname, "-relay-bin",
|
|
|
|
1, buf);
|
|
|
|
/* We send the warning only at startup, not after every RESET SLAVE */
|
2013-01-11 01:03:43 +01:00
|
|
|
if (!opt_relay_logname && !opt_relaylog_index_name && !name_warning_sent &&
|
|
|
|
!opt_bootstrap)
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
User didn't give us info to name the relay log index file.
|
|
|
|
Picking `hostname`-relay-bin.index like we do, causes replication to
|
|
|
|
fail if this slave's hostname is changed later. So, we would like to
|
|
|
|
instead require a name. But as we don't want to break many existing
|
|
|
|
setups, we only give warning, not error.
|
|
|
|
*/
|
|
|
|
sql_print_warning("Neither --relay-log nor --relay-log-index were used;"
|
|
|
|
" so replication "
|
|
|
|
"may break when this MySQL server acts as a "
|
|
|
|
"slave and has his hostname changed!! Please "
|
2011-03-28 02:02:24 +02:00
|
|
|
"use '--log-basename=#' or '--relay-log=%s' to avoid "
|
2011-03-23 16:59:41 +01:00
|
|
|
"this problem.", ln);
|
2006-10-31 12:23:14 +01:00
|
|
|
name_warning_sent= 1;
|
|
|
|
}
|
2011-05-03 14:01:11 +02:00
|
|
|
|
2012-09-28 01:06:56 +02:00
|
|
|
/* For multimaster, add connection name to relay log filenames */
|
|
|
|
Master_info* mi= rli->mi;
|
|
|
|
char buf_relay_logname[FN_REFLEN], buf_relaylog_index_name_buff[FN_REFLEN];
|
|
|
|
char *buf_relaylog_index_name= opt_relaylog_index_name;
|
|
|
|
|
2012-10-03 00:44:54 +02:00
|
|
|
create_logfile_name_with_suffix(buf_relay_logname, sizeof(buf_relay_logname),
|
2012-10-01 01:30:44 +02:00
|
|
|
ln, 1, &mi->connection_name);
|
2012-09-28 01:06:56 +02:00
|
|
|
ln= buf_relay_logname;
|
|
|
|
|
|
|
|
if (opt_relaylog_index_name)
|
|
|
|
{
|
|
|
|
buf_relaylog_index_name= buf_relaylog_index_name_buff;
|
2012-10-03 00:44:54 +02:00
|
|
|
create_logfile_name_with_suffix(buf_relaylog_index_name_buff,
|
2012-09-28 01:06:56 +02:00
|
|
|
sizeof(buf_relaylog_index_name_buff),
|
2012-10-01 01:30:44 +02:00
|
|
|
opt_relaylog_index_name, 0,
|
2012-09-28 01:06:56 +02:00
|
|
|
&mi->connection_name);
|
|
|
|
}
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
/*
|
|
|
|
note, that if open() fails, we'll still have index file open
|
|
|
|
but a destructor will take care of that
|
|
|
|
*/
|
2012-09-28 01:06:56 +02:00
|
|
|
if (rli->relay_log.open_index_file(buf_relaylog_index_name, ln, TRUE) ||
|
2012-09-13 14:31:29 +02:00
|
|
|
rli->relay_log.open(ln, LOG_BIN, 0, SEQ_READ_APPEND,
|
2012-10-01 01:30:44 +02:00
|
|
|
mi->rli.max_relay_log_size, 1, TRUE))
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2012-09-28 01:06:56 +02:00
|
|
|
sql_print_error("Failed when trying to open logs for '%s' in init_relay_log_info(). Error: %M", ln, my_errno);
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* if file does not exist */
|
|
|
|
if (access(fname,F_OK))
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
If someone removed the file from underneath our feet, just close
|
|
|
|
the old descriptor and re-create the old file
|
|
|
|
*/
|
|
|
|
if (info_fd >= 0)
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_file_close(info_fd, MYF(MY_WME));
|
|
|
|
if ((info_fd= mysql_file_open(key_file_relay_log_info,
|
|
|
|
fname, O_CREAT|O_RDWR|O_BINARY, MYF(MY_WME))) < 0)
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
sql_print_error("Failed to create a new relay log info file (\
|
|
|
|
file '%s', errno %d)", fname, my_errno);
|
2009-09-10 11:18:29 +02:00
|
|
|
msg= current_thd->stmt_da->message();
|
2006-10-31 12:23:14 +01:00
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
if (init_io_cache(&rli->info_file, info_fd, IO_SIZE*2, READ_CACHE, 0L,0,
|
|
|
|
MYF(MY_WME)))
|
|
|
|
{
|
|
|
|
sql_print_error("Failed to create a cache on relay log info file '%s'",
|
|
|
|
fname);
|
2009-09-10 11:18:29 +02:00
|
|
|
msg= current_thd->stmt_da->message();
|
2006-10-31 12:23:14 +01:00
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Init relay log with first entry in the relay index file */
|
|
|
|
if (init_relay_log_pos(rli,NullS,BIN_LOG_HEADER_SIZE,0 /* no data lock */,
|
|
|
|
&msg, 0))
|
|
|
|
{
|
|
|
|
sql_print_error("Failed to open the relay log 'FIRST' (relay_log_pos 4)");
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
rli->group_master_log_name[0]= 0;
|
|
|
|
rli->group_master_log_pos= 0;
|
|
|
|
rli->info_fd= info_fd;
|
|
|
|
}
|
|
|
|
else // file exists
|
|
|
|
{
|
|
|
|
if (info_fd >= 0)
|
|
|
|
reinit_io_cache(&rli->info_file, READ_CACHE, 0L,0,0);
|
|
|
|
else
|
|
|
|
{
|
|
|
|
int error=0;
|
2010-01-07 06:42:07 +01:00
|
|
|
if ((info_fd= mysql_file_open(key_file_relay_log_info,
|
|
|
|
fname, O_RDWR|O_BINARY, MYF(MY_WME))) < 0)
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
sql_print_error("\
|
|
|
|
Failed to open the existing relay log info file '%s' (errno %d)",
|
|
|
|
fname, my_errno);
|
|
|
|
error= 1;
|
|
|
|
}
|
|
|
|
else if (init_io_cache(&rli->info_file, info_fd,
|
|
|
|
IO_SIZE*2, READ_CACHE, 0L, 0, MYF(MY_WME)))
|
|
|
|
{
|
|
|
|
sql_print_error("Failed to create a cache on relay log info file '%s'",
|
|
|
|
fname);
|
|
|
|
error= 1;
|
|
|
|
}
|
|
|
|
if (error)
|
|
|
|
{
|
|
|
|
if (info_fd >= 0)
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_file_close(info_fd, MYF(0));
|
2006-10-31 12:23:14 +01:00
|
|
|
rli->info_fd= -1;
|
|
|
|
rli->relay_log.close(LOG_CLOSE_INDEX | LOG_CLOSE_STOP_EVENT);
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
rli->info_fd = info_fd;
|
|
|
|
int relay_log_pos, master_log_pos;
|
|
|
|
if (init_strvar_from_file(rli->group_relay_log_name,
|
|
|
|
sizeof(rli->group_relay_log_name),
|
|
|
|
&rli->info_file, "") ||
|
|
|
|
init_intvar_from_file(&relay_log_pos,
|
|
|
|
&rli->info_file, BIN_LOG_HEADER_SIZE) ||
|
|
|
|
init_strvar_from_file(rli->group_master_log_name,
|
|
|
|
sizeof(rli->group_master_log_name),
|
|
|
|
&rli->info_file, "") ||
|
|
|
|
init_intvar_from_file(&master_log_pos, &rli->info_file, 0))
|
|
|
|
{
|
|
|
|
msg="Error reading slave log configuration";
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
strmake(rli->event_relay_log_name,rli->group_relay_log_name,
|
|
|
|
sizeof(rli->event_relay_log_name)-1);
|
|
|
|
rli->group_relay_log_pos= rli->event_relay_log_pos= relay_log_pos;
|
|
|
|
rli->group_master_log_pos= master_log_pos;
|
|
|
|
|
2009-09-30 23:41:05 +02:00
|
|
|
if (rli->is_relay_log_recovery && init_recovery(rli->mi, &msg))
|
|
|
|
goto err;
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
if (init_relay_log_pos(rli,
|
|
|
|
rli->group_relay_log_name,
|
|
|
|
rli->group_relay_log_pos,
|
|
|
|
0 /* no data lock*/,
|
|
|
|
&msg, 0))
|
|
|
|
{
|
|
|
|
char llbuf[22];
|
|
|
|
sql_print_error("Failed to open the relay log '%s' (relay_log_pos %s)",
|
|
|
|
rli->group_relay_log_name,
|
|
|
|
llstr(rli->group_relay_log_pos, llbuf));
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifndef DBUG_OFF
|
|
|
|
{
|
|
|
|
char llbuf1[22], llbuf2[22];
|
|
|
|
DBUG_PRINT("info", ("my_b_tell(rli->cur_log)=%s rli->event_relay_log_pos=%s",
|
|
|
|
llstr(my_b_tell(rli->cur_log),llbuf1),
|
|
|
|
llstr(rli->event_relay_log_pos,llbuf2)));
|
|
|
|
DBUG_ASSERT(rli->event_relay_log_pos >= BIN_LOG_HEADER_SIZE);
|
|
|
|
DBUG_ASSERT(my_b_tell(rli->cur_log) == rli->event_relay_log_pos);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/*
|
|
|
|
Now change the cache from READ to WRITE - must do this
|
|
|
|
before flush_relay_log_info
|
|
|
|
*/
|
|
|
|
reinit_io_cache(&rli->info_file, WRITE_CACHE,0L,0,1);
|
|
|
|
if ((error= flush_relay_log_info(rli)))
|
2009-09-30 23:41:05 +02:00
|
|
|
{
|
|
|
|
msg= "Failed to flush relay log info file";
|
|
|
|
goto err;
|
|
|
|
}
|
2006-10-31 12:23:14 +01:00
|
|
|
if (count_relay_log_space(rli))
|
|
|
|
{
|
|
|
|
msg="Error counting relay log space";
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
rli->inited= 1;
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_RETURN(error);
|
|
|
|
|
|
|
|
err:
|
2009-09-23 15:21:29 +02:00
|
|
|
sql_print_error("%s", msg);
|
2006-10-31 12:23:14 +01:00
|
|
|
end_io_cache(&rli->info_file);
|
|
|
|
if (info_fd >= 0)
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_file_close(info_fd, MYF(0));
|
2006-10-31 12:23:14 +01:00
|
|
|
rli->info_fd= -1;
|
|
|
|
rli->relay_log.close(LOG_CLOSE_INDEX | LOG_CLOSE_STOP_EVENT);
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
static inline int add_relay_log(Relay_log_info* rli,LOG_INFO* linfo)
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
MY_STAT s;
|
|
|
|
DBUG_ENTER("add_relay_log");
|
2011-03-01 17:39:28 +01:00
|
|
|
if (!mysql_file_stat(key_file_relaylog,
|
2010-01-07 06:42:07 +01:00
|
|
|
linfo->log_file_name, &s, MYF(0)))
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
sql_print_error("log %s listed in the index, but failed to stat",
|
|
|
|
linfo->log_file_name);
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
rli->log_space_total += s.st_size;
|
|
|
|
#ifndef DBUG_OFF
|
|
|
|
char buf[22];
|
|
|
|
DBUG_PRINT("info",("log_space_total: %s", llstr(rli->log_space_total,buf)));
|
|
|
|
#endif
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
static int count_relay_log_space(Relay_log_info* rli)
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
LOG_INFO linfo;
|
|
|
|
DBUG_ENTER("count_relay_log_space");
|
|
|
|
rli->log_space_total= 0;
|
|
|
|
if (rli->relay_log.find_log_pos(&linfo, NullS, 1))
|
|
|
|
{
|
|
|
|
sql_print_error("Could not find first log while counting relay log space");
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
do
|
|
|
|
{
|
|
|
|
if (add_relay_log(rli,&linfo))
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
} while (!rli->relay_log.find_next_log(&linfo, 1));
|
|
|
|
/*
|
|
|
|
As we have counted everything, including what may have written in a
|
|
|
|
preceding write, we must reset bytes_written, or we may count some space
|
|
|
|
twice.
|
|
|
|
*/
|
|
|
|
rli->relay_log.reset_bytes_written();
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
2007-08-16 07:37:50 +02:00
|
|
|
Reset UNTIL condition for Relay_log_info
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
SYNOPSYS
|
|
|
|
clear_until_condition()
|
2007-08-16 07:37:50 +02:00
|
|
|
rli - Relay_log_info structure where UNTIL condition should be reset
|
2006-10-31 12:23:14 +01:00
|
|
|
*/
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
void Relay_log_info::clear_until_condition()
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
DBUG_ENTER("clear_until_condition");
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
until_condition= Relay_log_info::UNTIL_NONE;
|
2006-10-31 12:23:14 +01:00
|
|
|
until_log_name[0]= 0;
|
|
|
|
until_log_pos= 0;
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Open the given relay log
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
init_relay_log_pos()
|
|
|
|
rli Relay information (will be initialized)
|
|
|
|
log Name of relay log file to read from. NULL = First log
|
|
|
|
pos Position in relay log file
|
|
|
|
need_data_lock Set to 1 if this functions should do mutex locks
|
|
|
|
errmsg Store pointer to error message here
|
|
|
|
look_for_description_event
|
|
|
|
1 if we should look for such an event. We only need
|
|
|
|
this when the SQL thread starts and opens an existing
|
|
|
|
relay log and has to execute it (possibly from an
|
|
|
|
offset >4); then we need to read the first event of
|
|
|
|
the relay log to be able to parse the events we have
|
|
|
|
to execute.
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
- Close old open relay log files.
|
|
|
|
- If we are using the same relay log as the running IO-thread, then set
|
|
|
|
rli->cur_log to point to the same IO_CACHE entry.
|
|
|
|
- If not, open the 'log' binary file.
|
|
|
|
|
|
|
|
TODO
|
|
|
|
- check proper initialization of group_master_log_name/group_master_log_pos
|
|
|
|
|
|
|
|
RETURN VALUES
|
|
|
|
0 ok
|
|
|
|
1 error. errmsg is set to point to the error message
|
|
|
|
*/
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
int init_relay_log_pos(Relay_log_info* rli,const char* log,
|
2006-10-31 12:23:14 +01:00
|
|
|
ulonglong pos, bool need_data_lock,
|
|
|
|
const char** errmsg,
|
|
|
|
bool look_for_description_event)
|
|
|
|
{
|
|
|
|
DBUG_ENTER("init_relay_log_pos");
|
2006-11-27 00:47:38 +01:00
|
|
|
DBUG_PRINT("info", ("pos: %lu", (ulong) pos));
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
*errmsg=0;
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_t *log_lock= rli->relay_log.get_log_lock();
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
if (need_data_lock)
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_lock(&rli->data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
Slave threads are not the only users of init_relay_log_pos(). CHANGE MASTER
|
|
|
|
is, too, and init_slave() too; these 2 functions allocate a description
|
|
|
|
event in init_relay_log_pos, which is not freed by the terminating SQL slave
|
|
|
|
thread as that thread is not started by these functions. So we have to free
|
|
|
|
the description_event here, in case, so that there is no memory leak in
|
|
|
|
running, say, CHANGE MASTER.
|
|
|
|
*/
|
|
|
|
delete rli->relay_log.description_event_for_exec;
|
|
|
|
/*
|
|
|
|
By default the relay log is in binlog format 3 (4.0).
|
|
|
|
Even if format is 4, this will work enough to read the first event
|
|
|
|
(Format_desc) (remember that format 4 is just lenghtened compared to format
|
|
|
|
3; format 3 is a prefix of format 4).
|
|
|
|
*/
|
|
|
|
rli->relay_log.description_event_for_exec= new
|
|
|
|
Format_description_log_event(3);
|
|
|
|
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_lock(log_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
/* Close log file and free buffers if it's already open */
|
|
|
|
if (rli->cur_log_fd >= 0)
|
|
|
|
{
|
|
|
|
end_io_cache(&rli->cache_buf);
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_file_close(rli->cur_log_fd, MYF(MY_WME));
|
2006-10-31 12:23:14 +01:00
|
|
|
rli->cur_log_fd = -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
rli->group_relay_log_pos = rli->event_relay_log_pos = pos;
|
|
|
|
|
|
|
|
/*
|
|
|
|
Test to see if the previous run was with the skip of purging
|
|
|
|
If yes, we do not purge when we restart
|
|
|
|
*/
|
|
|
|
if (rli->relay_log.find_log_pos(&rli->linfo, NullS, 1))
|
|
|
|
{
|
|
|
|
*errmsg="Could not find first log during relay log initialization";
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (log && rli->relay_log.find_log_pos(&rli->linfo, log, 1))
|
|
|
|
{
|
|
|
|
*errmsg="Could not find target log during relay log initialization";
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
strmake(rli->group_relay_log_name,rli->linfo.log_file_name,
|
|
|
|
sizeof(rli->group_relay_log_name)-1);
|
|
|
|
strmake(rli->event_relay_log_name,rli->linfo.log_file_name,
|
|
|
|
sizeof(rli->event_relay_log_name)-1);
|
|
|
|
if (rli->relay_log.is_active(rli->linfo.log_file_name))
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
The IO thread is using this log file.
|
|
|
|
In this case, we will use the same IO_CACHE pointer to
|
|
|
|
read data as the IO thread is using to write data.
|
|
|
|
*/
|
|
|
|
my_b_seek((rli->cur_log=rli->relay_log.get_log_file()), (off_t)0);
|
|
|
|
if (check_binlog_magic(rli->cur_log,errmsg))
|
|
|
|
goto err;
|
|
|
|
rli->cur_log_old_open_count=rli->relay_log.get_open_count();
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
Open the relay log and set rli->cur_log to point at this one
|
|
|
|
*/
|
|
|
|
if ((rli->cur_log_fd=open_binlog(&rli->cache_buf,
|
|
|
|
rli->linfo.log_file_name,errmsg)) < 0)
|
|
|
|
goto err;
|
|
|
|
rli->cur_log = &rli->cache_buf;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
In all cases, check_binlog_magic() has been called so we're at offset 4 for
|
|
|
|
sure.
|
|
|
|
*/
|
|
|
|
if (pos > BIN_LOG_HEADER_SIZE) /* If pos<=4, we stay at 4 */
|
|
|
|
{
|
|
|
|
Log_event* ev;
|
|
|
|
while (look_for_description_event)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
Read the possible Format_description_log_event; if position
|
|
|
|
was 4, no need, it will be read naturally.
|
|
|
|
*/
|
|
|
|
DBUG_PRINT("info",("looking for a Format_description_log_event"));
|
|
|
|
|
|
|
|
if (my_b_tell(rli->cur_log) >= pos)
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
Because of we have rli->data_lock and log_lock, we can safely read an
|
|
|
|
event
|
|
|
|
*/
|
2011-05-03 14:01:11 +02:00
|
|
|
if (!(ev= Log_event::read_log_event(rli->cur_log, 0,
|
|
|
|
rli->relay_log.description_event_for_exec,
|
|
|
|
opt_slave_sql_verify_checksum)))
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
DBUG_PRINT("info",("could not read event, rli->cur_log->error=%d",
|
|
|
|
rli->cur_log->error));
|
|
|
|
if (rli->cur_log->error) /* not EOF */
|
|
|
|
{
|
|
|
|
*errmsg= "I/O error reading event at position 4";
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
else if (ev->get_type_code() == FORMAT_DESCRIPTION_EVENT)
|
|
|
|
{
|
|
|
|
DBUG_PRINT("info",("found Format_description_log_event"));
|
|
|
|
delete rli->relay_log.description_event_for_exec;
|
|
|
|
rli->relay_log.description_event_for_exec= (Format_description_log_event*) ev;
|
|
|
|
/*
|
|
|
|
As ev was returned by read_log_event, it has passed is_valid(), so
|
|
|
|
my_malloc() in ctor worked, no need to check again.
|
|
|
|
*/
|
|
|
|
/*
|
|
|
|
Ok, we found a Format_description event. But it is not sure that this
|
|
|
|
describes the whole relay log; indeed, one can have this sequence
|
|
|
|
(starting from position 4):
|
|
|
|
Format_desc (of slave)
|
|
|
|
Rotate (of master)
|
|
|
|
Format_desc (of master)
|
|
|
|
So the Format_desc which really describes the rest of the relay log
|
|
|
|
is the 3rd event (it can't be further than that, because we rotate
|
|
|
|
the relay log when we queue a Rotate event from the master).
|
|
|
|
But what describes the Rotate is the first Format_desc.
|
|
|
|
So what we do is:
|
|
|
|
go on searching for Format_description events, until you exceed the
|
|
|
|
position (argument 'pos') or until you find another event than Rotate
|
|
|
|
or Format_desc.
|
|
|
|
*/
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
DBUG_PRINT("info",("found event of another type=%d",
|
|
|
|
ev->get_type_code()));
|
|
|
|
look_for_description_event= (ev->get_type_code() == ROTATE_EVENT);
|
|
|
|
delete ev;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
my_b_seek(rli->cur_log,(off_t)pos);
|
|
|
|
#ifndef DBUG_OFF
|
|
|
|
{
|
|
|
|
char llbuf1[22], llbuf2[22];
|
|
|
|
DBUG_PRINT("info", ("my_b_tell(rli->cur_log)=%s rli->event_relay_log_pos=%s",
|
|
|
|
llstr(my_b_tell(rli->cur_log),llbuf1),
|
|
|
|
llstr(rli->event_relay_log_pos,llbuf2)));
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
err:
|
|
|
|
/*
|
|
|
|
If we don't purge, we can't honour relay_log_space_limit ;
|
|
|
|
silently discard it
|
|
|
|
*/
|
|
|
|
if (!relay_log_purge)
|
|
|
|
rli->log_space_limit= 0;
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_cond_broadcast(&rli->data_cond);
|
2006-10-31 12:23:14 +01:00
|
|
|
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(log_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
if (need_data_lock)
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
if (!rli->relay_log.description_event_for_exec->is_valid() && !*errmsg)
|
|
|
|
*errmsg= "Invalid Format_description log event; could be out of memory";
|
|
|
|
|
|
|
|
DBUG_RETURN ((*errmsg) ? 1 : 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Waits until the SQL thread reaches (has executed up to) the
|
|
|
|
log/position or timed out.
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
wait_for_pos()
|
|
|
|
thd client thread that sent SELECT MASTER_POS_WAIT
|
|
|
|
log_name log name to wait for
|
|
|
|
log_pos position to wait for
|
|
|
|
timeout timeout in seconds before giving up waiting
|
|
|
|
|
|
|
|
NOTES
|
|
|
|
timeout is longlong whereas it should be ulong ; but this is
|
|
|
|
to catch if the user submitted a negative timeout.
|
|
|
|
|
|
|
|
RETURN VALUES
|
|
|
|
-2 improper arguments (log_pos<0)
|
|
|
|
or slave not running, or master info changed
|
|
|
|
during the function's execution,
|
|
|
|
or client thread killed. -2 is translated to NULL by caller
|
|
|
|
-1 timed out
|
|
|
|
>=0 number of log events the function had to wait
|
|
|
|
before reaching the desired log/position
|
|
|
|
*/
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
int Relay_log_info::wait_for_pos(THD* thd, String* log_name,
|
2006-10-31 12:23:14 +01:00
|
|
|
longlong log_pos,
|
|
|
|
longlong timeout)
|
|
|
|
{
|
|
|
|
int event_count = 0;
|
|
|
|
ulong init_abort_pos_wait;
|
|
|
|
int error=0;
|
|
|
|
struct timespec abstime; // for timeout checking
|
fixes for test failures
and small collateral changes
mysql-test/lib/My/Test.pm:
somehow with "print" we get truncated writes sometimes
mysql-test/suite/perfschema/r/digest_table_full.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/dml_handler.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/information_schema.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/nesting.result:
this differs, because we don't rewrite general log queries, and multi-statement
packets are logged as a one entry. this result file is identical to what mysql-5.6.5
produces with the --log-raw option.
mysql-test/suite/perfschema/r/relaylog.result:
MariaDB modifies the binlog index file directly, while MySQL 5.6 has a feature "crash-safe binlog index" and modifies a special "crash-safe" shadow copy of the index file and then moves it over. That's why this test shows "NONE" index file writes in MySQL and "MANY" in MariaDB.
mysql-test/suite/perfschema/r/server_init.result:
MariaDB initializes the "manager" resources from the "manager" thread, and starts this thread only when --flush-time is not 0. MySQL 5.6 initializes "manager" resources unconditionally on server startup.
mysql-test/suite/perfschema/r/stage_mdl_global.result:
this differs, because MariaDB disables query cache when query_cache_size=0. MySQL does not
do that, and this causes useless mutex locks and waits.
mysql-test/suite/perfschema/r/statement_digest.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_consumers.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_long_query.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/rpl/r/rpl_mixed_drop_create_temp_table.result:
will be updated to match 5.6 when alfranio.correia@oracle.com-20110512172919-c1b5kmum4h52g0ni and anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y are merged
mysql-test/suite/rpl/r/rpl_non_direct_mixed_mixing_engines.result:
will be updated to match 5.6 when anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y is merged
2012-09-27 20:09:46 +02:00
|
|
|
PSI_stage_info old_stage;
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::wait_for_pos");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
if (!inited)
|
2008-03-14 17:52:57 +01:00
|
|
|
DBUG_RETURN(-2);
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
DBUG_PRINT("enter",("log_name: '%s' log_pos: %lu timeout: %lu",
|
|
|
|
log_name->c_ptr(), (ulong) log_pos, (ulong) timeout));
|
|
|
|
|
|
|
|
set_timespec(abstime,timeout);
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_lock(&data_lock);
|
fixes for test failures
and small collateral changes
mysql-test/lib/My/Test.pm:
somehow with "print" we get truncated writes sometimes
mysql-test/suite/perfschema/r/digest_table_full.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/dml_handler.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/information_schema.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/nesting.result:
this differs, because we don't rewrite general log queries, and multi-statement
packets are logged as a one entry. this result file is identical to what mysql-5.6.5
produces with the --log-raw option.
mysql-test/suite/perfschema/r/relaylog.result:
MariaDB modifies the binlog index file directly, while MySQL 5.6 has a feature "crash-safe binlog index" and modifies a special "crash-safe" shadow copy of the index file and then moves it over. That's why this test shows "NONE" index file writes in MySQL and "MANY" in MariaDB.
mysql-test/suite/perfschema/r/server_init.result:
MariaDB initializes the "manager" resources from the "manager" thread, and starts this thread only when --flush-time is not 0. MySQL 5.6 initializes "manager" resources unconditionally on server startup.
mysql-test/suite/perfschema/r/stage_mdl_global.result:
this differs, because MariaDB disables query cache when query_cache_size=0. MySQL does not
do that, and this causes useless mutex locks and waits.
mysql-test/suite/perfschema/r/statement_digest.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_consumers.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_long_query.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/rpl/r/rpl_mixed_drop_create_temp_table.result:
will be updated to match 5.6 when alfranio.correia@oracle.com-20110512172919-c1b5kmum4h52g0ni and anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y are merged
mysql-test/suite/rpl/r/rpl_non_direct_mixed_mixing_engines.result:
will be updated to match 5.6 when anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y is merged
2012-09-27 20:09:46 +02:00
|
|
|
thd->ENTER_COND(&data_cond, &data_lock,
|
|
|
|
&stage_waiting_for_the_slave_thread_to_advance_position,
|
|
|
|
&old_stage);
|
2006-10-31 12:23:14 +01:00
|
|
|
/*
|
|
|
|
This function will abort when it notices that some CHANGE MASTER or
|
|
|
|
RESET MASTER has changed the master info.
|
|
|
|
To catch this, these commands modify abort_pos_wait ; We just monitor
|
|
|
|
abort_pos_wait and see if it has changed.
|
|
|
|
Why do we have this mechanism instead of simply monitoring slave_running
|
|
|
|
in the loop (we do this too), as CHANGE MASTER/RESET SLAVE require that
|
|
|
|
the SQL thread be stopped?
|
|
|
|
This is becasue if someones does:
|
|
|
|
STOP SLAVE;CHANGE MASTER/RESET SLAVE; START SLAVE;
|
|
|
|
the change may happen very quickly and we may not notice that
|
|
|
|
slave_running briefly switches between 1/0/1.
|
|
|
|
*/
|
|
|
|
init_abort_pos_wait= abort_pos_wait;
|
|
|
|
|
|
|
|
/*
|
|
|
|
We'll need to
|
|
|
|
handle all possible log names comparisons (e.g. 999 vs 1000).
|
|
|
|
We use ulong for string->number conversion ; this is no
|
|
|
|
stronger limitation than in find_uniq_filename in sql/log.cc
|
|
|
|
*/
|
|
|
|
ulong log_name_extension;
|
|
|
|
char log_name_tmp[FN_REFLEN]; //make a char[] from String
|
|
|
|
|
|
|
|
strmake(log_name_tmp, log_name->ptr(), min(log_name->length(), FN_REFLEN-1));
|
|
|
|
|
|
|
|
char *p= fn_ext(log_name_tmp);
|
|
|
|
char *p_end;
|
|
|
|
if (!*p || log_pos<0)
|
|
|
|
{
|
|
|
|
error= -2; //means improper arguments
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
// Convert 0-3 to 4
|
|
|
|
log_pos= max(log_pos, BIN_LOG_HEADER_SIZE);
|
|
|
|
/* p points to '.' */
|
|
|
|
log_name_extension= strtoul(++p, &p_end, 10);
|
|
|
|
/*
|
|
|
|
p_end points to the first invalid character.
|
|
|
|
If it equals to p, no digits were found, error.
|
|
|
|
If it contains '\0' it means conversion went ok.
|
|
|
|
*/
|
|
|
|
if (p_end==p || *p_end)
|
|
|
|
{
|
|
|
|
error= -2;
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* The "compare and wait" main loop */
|
|
|
|
while (!thd->killed &&
|
|
|
|
init_abort_pos_wait == abort_pos_wait &&
|
|
|
|
slave_running)
|
|
|
|
{
|
|
|
|
bool pos_reached;
|
|
|
|
int cmp_result= 0;
|
|
|
|
|
|
|
|
DBUG_PRINT("info",
|
|
|
|
("init_abort_pos_wait: %ld abort_pos_wait: %ld",
|
|
|
|
init_abort_pos_wait, abort_pos_wait));
|
|
|
|
DBUG_PRINT("info",("group_master_log_name: '%s' pos: %lu",
|
|
|
|
group_master_log_name, (ulong) group_master_log_pos));
|
|
|
|
|
|
|
|
/*
|
|
|
|
group_master_log_name can be "", if we are just after a fresh
|
|
|
|
replication start or after a CHANGE MASTER TO MASTER_HOST/PORT
|
|
|
|
(before we have executed one Rotate event from the master) or
|
|
|
|
(rare) if the user is doing a weird slave setup (see next
|
|
|
|
paragraph). If group_master_log_name is "", we assume we don't
|
|
|
|
have enough info to do the comparison yet, so we just wait until
|
|
|
|
more data. In this case master_log_pos is always 0 except if
|
|
|
|
somebody (wrongly) sets this slave to be a slave of itself
|
|
|
|
without using --replicate-same-server-id (an unsupported
|
|
|
|
configuration which does nothing), then group_master_log_pos
|
|
|
|
will grow and group_master_log_name will stay "".
|
|
|
|
*/
|
|
|
|
if (*group_master_log_name)
|
|
|
|
{
|
|
|
|
char *basename= (group_master_log_name +
|
|
|
|
dirname_length(group_master_log_name));
|
|
|
|
/*
|
|
|
|
First compare the parts before the extension.
|
|
|
|
Find the dot in the master's log basename,
|
|
|
|
and protect against user's input error :
|
|
|
|
if the names do not match up to '.' included, return error
|
|
|
|
*/
|
|
|
|
char *q= (char*)(fn_ext(basename)+1);
|
|
|
|
if (strncmp(basename, log_name_tmp, (int)(q-basename)))
|
|
|
|
{
|
|
|
|
error= -2;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
// Now compare extensions.
|
|
|
|
char *q_end;
|
|
|
|
ulong group_master_log_name_extension= strtoul(q, &q_end, 10);
|
|
|
|
if (group_master_log_name_extension < log_name_extension)
|
|
|
|
cmp_result= -1 ;
|
|
|
|
else
|
|
|
|
cmp_result= (group_master_log_name_extension > log_name_extension) ? 1 : 0 ;
|
|
|
|
|
|
|
|
pos_reached= ((!cmp_result && group_master_log_pos >= (ulonglong)log_pos) ||
|
|
|
|
cmp_result > 0);
|
|
|
|
if (pos_reached || thd->killed)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
//wait for master update, with optional timeout.
|
|
|
|
|
|
|
|
DBUG_PRINT("info",("Waiting for master update"));
|
|
|
|
/*
|
2010-01-07 06:42:07 +01:00
|
|
|
We are going to mysql_cond_(timed)wait(); if the SQL thread stops it
|
2006-10-31 12:23:14 +01:00
|
|
|
will wake us up.
|
|
|
|
*/
|
2010-10-28 16:31:21 +02:00
|
|
|
thd_wait_begin(thd, THD_WAIT_BINLOG);
|
2006-10-31 12:23:14 +01:00
|
|
|
if (timeout > 0)
|
|
|
|
{
|
|
|
|
/*
|
2010-01-07 06:42:07 +01:00
|
|
|
Note that mysql_cond_timedwait checks for the timeout
|
2006-10-31 12:23:14 +01:00
|
|
|
before for the condition ; i.e. it returns ETIMEDOUT
|
|
|
|
if the system time equals or exceeds the time specified by abstime
|
|
|
|
before the condition variable is signaled or broadcast, _or_ if
|
|
|
|
the absolute time specified by abstime has already passed at the time
|
|
|
|
of the call.
|
2010-01-07 06:42:07 +01:00
|
|
|
For that reason, mysql_cond_timedwait will do the "timeoutting" job
|
2006-10-31 12:23:14 +01:00
|
|
|
even if its condition is always immediately signaled (case of a loaded
|
|
|
|
master).
|
|
|
|
*/
|
2010-01-07 06:42:07 +01:00
|
|
|
error= mysql_cond_timedwait(&data_cond, &data_lock, &abstime);
|
2006-10-31 12:23:14 +01:00
|
|
|
}
|
|
|
|
else
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_cond_wait(&data_cond, &data_lock);
|
2010-10-28 16:31:21 +02:00
|
|
|
thd_wait_end(thd);
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_PRINT("info",("Got signal of master update or timed out"));
|
|
|
|
if (error == ETIMEDOUT || error == ETIME)
|
|
|
|
{
|
|
|
|
error= -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
error=0;
|
|
|
|
event_count++;
|
|
|
|
DBUG_PRINT("info",("Testing if killed or SQL thread not running"));
|
|
|
|
}
|
|
|
|
|
|
|
|
err:
|
fixes for test failures
and small collateral changes
mysql-test/lib/My/Test.pm:
somehow with "print" we get truncated writes sometimes
mysql-test/suite/perfschema/r/digest_table_full.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/dml_handler.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/information_schema.result:
host table is not ported over yet
mysql-test/suite/perfschema/r/nesting.result:
this differs, because we don't rewrite general log queries, and multi-statement
packets are logged as a one entry. this result file is identical to what mysql-5.6.5
produces with the --log-raw option.
mysql-test/suite/perfschema/r/relaylog.result:
MariaDB modifies the binlog index file directly, while MySQL 5.6 has a feature "crash-safe binlog index" and modifies a special "crash-safe" shadow copy of the index file and then moves it over. That's why this test shows "NONE" index file writes in MySQL and "MANY" in MariaDB.
mysql-test/suite/perfschema/r/server_init.result:
MariaDB initializes the "manager" resources from the "manager" thread, and starts this thread only when --flush-time is not 0. MySQL 5.6 initializes "manager" resources unconditionally on server startup.
mysql-test/suite/perfschema/r/stage_mdl_global.result:
this differs, because MariaDB disables query cache when query_cache_size=0. MySQL does not
do that, and this causes useless mutex locks and waits.
mysql-test/suite/perfschema/r/statement_digest.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_consumers.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/perfschema/r/statement_digest_long_query.result:
md5 hashes of statement digests differ, because yacc token codes are different in mariadb
mysql-test/suite/rpl/r/rpl_mixed_drop_create_temp_table.result:
will be updated to match 5.6 when alfranio.correia@oracle.com-20110512172919-c1b5kmum4h52g0ni and anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y are merged
mysql-test/suite/rpl/r/rpl_non_direct_mixed_mixing_engines.result:
will be updated to match 5.6 when anders.song@greatopensource.com-20110105052107-zoab0bsf5a6xxk2y is merged
2012-09-27 20:09:46 +02:00
|
|
|
thd->EXIT_COND(&old_stage);
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_PRINT("exit",("killed: %d abort: %d slave_running: %d \
|
|
|
|
improper_arguments: %d timed_out: %d",
|
|
|
|
thd->killed_errno(),
|
|
|
|
(int) (init_abort_pos_wait != abort_pos_wait),
|
|
|
|
(int) slave_running,
|
|
|
|
(int) (error == -2),
|
|
|
|
(int) (error == -1)));
|
|
|
|
if (thd->killed || init_abort_pos_wait != abort_pos_wait ||
|
|
|
|
!slave_running)
|
|
|
|
{
|
|
|
|
error= -2;
|
|
|
|
}
|
|
|
|
DBUG_RETURN( error ? error : event_count );
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
void Relay_log_info::inc_group_relay_log_pos(ulonglong log_pos,
|
2006-10-31 12:23:14 +01:00
|
|
|
bool skip_lock)
|
|
|
|
{
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::inc_group_relay_log_pos");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
if (!skip_lock)
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_lock(&data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
inc_event_relay_log_pos();
|
|
|
|
group_relay_log_pos= event_relay_log_pos;
|
|
|
|
strmake(group_relay_log_name,event_relay_log_name,
|
|
|
|
sizeof(group_relay_log_name)-1);
|
|
|
|
|
|
|
|
notify_group_relay_log_name_update();
|
|
|
|
|
|
|
|
/*
|
|
|
|
If the slave does not support transactions and replicates a transaction,
|
|
|
|
users should not trust group_master_log_pos (which they can display with
|
|
|
|
SHOW SLAVE STATUS or read from relay-log.info), because to compute
|
|
|
|
group_master_log_pos the slave relies on log_pos stored in the master's
|
|
|
|
binlog, but if we are in a master's transaction these positions are always
|
|
|
|
the BEGIN's one (excepted for the COMMIT), so group_master_log_pos does
|
|
|
|
not advance as it should on the non-transactional slave (it advances by
|
|
|
|
big leaps, whereas it should advance by small leaps).
|
|
|
|
*/
|
|
|
|
/*
|
|
|
|
In 4.x we used the event's len to compute the positions here. This is
|
|
|
|
wrong if the event was 3.23/4.0 and has been converted to 5.0, because
|
|
|
|
then the event's len is not what is was in the master's binlog, so this
|
|
|
|
will make a wrong group_master_log_pos (yes it's a bug in 3.23->4.0
|
|
|
|
replication: Exec_master_log_pos is wrong). Only way to solve this is to
|
|
|
|
have the original offset of the end of the event the relay log. This is
|
|
|
|
what we do in 5.0: log_pos has become "end_log_pos" (because the real use
|
|
|
|
of log_pos in 4.0 was to compute the end_log_pos; so better to store
|
|
|
|
end_log_pos instead of begin_log_pos.
|
|
|
|
If we had not done this fix here, the problem would also have appeared
|
|
|
|
when the slave and master are 5.0 but with different event length (for
|
|
|
|
example the slave is more recent than the master and features the event
|
|
|
|
UID). It would give false MASTER_POS_WAIT, false Exec_master_log_pos in
|
|
|
|
SHOW SLAVE STATUS, and so the user would do some CHANGE MASTER using this
|
|
|
|
value which would lead to badly broken replication.
|
|
|
|
Even the relay_log_pos will be corrupted in this case, because the len is
|
|
|
|
the relay log is not "val".
|
|
|
|
With the end_log_pos solution, we avoid computations involving lengthes.
|
|
|
|
*/
|
|
|
|
DBUG_PRINT("info", ("log_pos: %lu group_master_log_pos: %lu",
|
|
|
|
(long) log_pos, (long) group_master_log_pos));
|
|
|
|
if (log_pos) // 3.23 binlogs don't have log_posx
|
|
|
|
{
|
|
|
|
group_master_log_pos= log_pos;
|
|
|
|
}
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_cond_broadcast(&data_cond);
|
2006-10-31 12:23:14 +01:00
|
|
|
if (!skip_lock)
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
void Relay_log_info::close_temporary_tables()
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
TABLE *table,*next;
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::close_temporary_tables");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
for (table=save_temporary_tables ; table ; table=next)
|
|
|
|
{
|
|
|
|
next=table->next;
|
|
|
|
/*
|
|
|
|
Don't ask for disk deletion. For now, anyway they will be deleted when
|
|
|
|
slave restarts, but it is a better intention to not delete them.
|
|
|
|
*/
|
2006-11-27 00:47:38 +01:00
|
|
|
DBUG_PRINT("info", ("table: 0x%lx", (long) table));
|
2006-10-31 12:23:14 +01:00
|
|
|
close_temporary(table, 1, 0);
|
|
|
|
}
|
|
|
|
save_temporary_tables= 0;
|
|
|
|
slave_open_temp_tables= 0;
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
purge_relay_logs()
|
|
|
|
|
|
|
|
NOTES
|
|
|
|
Assumes to have a run lock on rli and that no slave thread are running.
|
|
|
|
*/
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
int purge_relay_logs(Relay_log_info* rli, THD *thd, bool just_reset,
|
2006-10-31 12:23:14 +01:00
|
|
|
const char** errmsg)
|
|
|
|
{
|
|
|
|
int error=0;
|
|
|
|
DBUG_ENTER("purge_relay_logs");
|
|
|
|
|
|
|
|
/*
|
|
|
|
Even if rli->inited==0, we still try to empty rli->master_log_* variables.
|
|
|
|
Indeed, rli->inited==0 does not imply that they already are empty.
|
|
|
|
It could be that slave's info initialization partly succeeded :
|
|
|
|
for example if relay-log.info existed but *relay-bin*.*
|
|
|
|
have been manually removed, init_relay_log_info reads the old
|
|
|
|
relay-log.info and fills rli->master_log_*, then init_relay_log_info
|
|
|
|
checks for the existence of the relay log, this fails and
|
|
|
|
init_relay_log_info leaves rli->inited to 0.
|
|
|
|
In that pathological case, rli->master_log_pos* will be properly reinited
|
|
|
|
at the next START SLAVE (as RESET SLAVE or CHANGE
|
|
|
|
MASTER, the callers of purge_relay_logs, will delete bogus *.info files
|
|
|
|
or replace them with correct files), however if the user does SHOW SLAVE
|
|
|
|
STATUS before START SLAVE, he will see old, confusing rli->master_log_*.
|
|
|
|
In other words, we reinit rli->master_log_* for SHOW SLAVE STATUS
|
|
|
|
to display fine in any case.
|
|
|
|
*/
|
|
|
|
|
|
|
|
rli->group_master_log_name[0]= 0;
|
|
|
|
rli->group_master_log_pos= 0;
|
|
|
|
|
|
|
|
if (!rli->inited)
|
|
|
|
{
|
|
|
|
DBUG_PRINT("info", ("rli->inited == 0"));
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
DBUG_ASSERT(rli->slave_running == 0);
|
|
|
|
DBUG_ASSERT(rli->mi->slave_running == 0);
|
|
|
|
|
|
|
|
rli->slave_skip_counter=0;
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_lock(&rli->data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
we close the relay log fd possibly left open by the slave SQL thread,
|
|
|
|
to be able to delete it; the relay log fd possibly left open by the slave
|
|
|
|
I/O thread will be closed naturally in reset_logs() by the
|
|
|
|
close(LOG_CLOSE_TO_BE_OPENED) call
|
|
|
|
*/
|
|
|
|
if (rli->cur_log_fd >= 0)
|
|
|
|
{
|
|
|
|
end_io_cache(&rli->cache_buf);
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_file_close(rli->cur_log_fd, MYF(MY_WME));
|
2006-10-31 12:23:14 +01:00
|
|
|
rli->cur_log_fd= -1;
|
|
|
|
}
|
|
|
|
|
2012-09-28 20:22:24 +02:00
|
|
|
if (rli->relay_log.reset_logs(thd, !just_reset))
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
*errmsg = "Failed during log reset";
|
|
|
|
error=1;
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
if (!just_reset)
|
2012-09-28 20:22:24 +02:00
|
|
|
{
|
|
|
|
/* Save name of used relay log file */
|
|
|
|
strmake(rli->group_relay_log_name, rli->relay_log.get_log_fname(),
|
|
|
|
sizeof(rli->group_relay_log_name)-1);
|
|
|
|
strmake(rli->event_relay_log_name, rli->relay_log.get_log_fname(),
|
|
|
|
sizeof(rli->event_relay_log_name)-1);
|
|
|
|
rli->group_relay_log_pos= rli->event_relay_log_pos= BIN_LOG_HEADER_SIZE;
|
|
|
|
rli->log_space_total= 0;
|
|
|
|
|
|
|
|
if (count_relay_log_space(rli))
|
|
|
|
{
|
|
|
|
*errmsg= "Error counting relay log space";
|
|
|
|
error=1;
|
|
|
|
goto err;
|
|
|
|
}
|
2006-10-31 12:23:14 +01:00
|
|
|
error= init_relay_log_pos(rli, rli->group_relay_log_name,
|
|
|
|
rli->group_relay_log_pos,
|
|
|
|
0 /* do not need data lock */, errmsg, 0);
|
2012-09-28 20:22:24 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
/* Ensure relay log names are not used */
|
|
|
|
rli->group_relay_log_name[0]= rli->event_relay_log_name[0]= 0;
|
|
|
|
}
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
err:
|
|
|
|
#ifndef DBUG_OFF
|
|
|
|
char buf[22];
|
|
|
|
#endif
|
|
|
|
DBUG_PRINT("info",("log_space_total: %s",llstr(rli->log_space_total,buf)));
|
2010-01-07 06:42:07 +01:00
|
|
|
mysql_mutex_unlock(&rli->data_lock);
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_RETURN(error);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Check if condition stated in UNTIL clause of START SLAVE is reached.
|
|
|
|
SYNOPSYS
|
2007-08-16 07:37:50 +02:00
|
|
|
Relay_log_info::is_until_satisfied()
|
2008-02-27 18:46:06 +01:00
|
|
|
master_beg_pos position of the beginning of to be executed event
|
|
|
|
(not log_pos member of the event that points to the
|
|
|
|
beginning of the following event)
|
|
|
|
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
DESCRIPTION
|
|
|
|
Checks if UNTIL condition is reached. Uses caching result of last
|
|
|
|
comparison of current log file name and target log file name. So cached
|
|
|
|
value should be invalidated if current log file name changes
|
2007-08-16 07:37:50 +02:00
|
|
|
(see Relay_log_info::notify_... functions).
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
This caching is needed to avoid of expensive string comparisons and
|
|
|
|
strtol() conversions needed for log names comparison. We don't need to
|
|
|
|
compare them each time this function is called, we only need to do this
|
|
|
|
when current log name changes. If we have UNTIL_MASTER_POS condition we
|
2007-03-30 15:29:30 +02:00
|
|
|
need to do this only after Rotate_log_event::do_apply_event() (which is
|
2006-10-31 12:23:14 +01:00
|
|
|
rare, so caching gives real benifit), and if we have UNTIL_RELAY_POS
|
|
|
|
condition then we should invalidate cached comarison value after
|
|
|
|
inc_group_relay_log_pos() which called for each group of events (so we
|
|
|
|
have some benefit if we have something like queries that use
|
|
|
|
autoincrement or if we have transactions).
|
|
|
|
|
|
|
|
Should be called ONLY if until_condition != UNTIL_NONE !
|
|
|
|
RETURN VALUE
|
|
|
|
true - condition met or error happened (condition seems to have
|
|
|
|
bad log file name)
|
|
|
|
false - condition not met
|
|
|
|
*/
|
|
|
|
|
2009-12-14 17:32:22 +01:00
|
|
|
bool Relay_log_info::is_until_satisfied(THD *thd, Log_event *ev)
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
|
|
|
const char *log_name;
|
|
|
|
ulonglong log_pos;
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::is_until_satisfied");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
DBUG_ASSERT(until_condition != UNTIL_NONE);
|
|
|
|
|
|
|
|
if (until_condition == UNTIL_MASTER_POS)
|
|
|
|
{
|
2009-12-14 17:32:22 +01:00
|
|
|
if (ev && ev->server_id == (uint32) ::server_id && !replicate_same_server_id)
|
2009-12-14 17:50:22 +01:00
|
|
|
DBUG_RETURN(FALSE);
|
2006-10-31 12:23:14 +01:00
|
|
|
log_name= group_master_log_name;
|
2009-12-14 17:32:22 +01:00
|
|
|
log_pos= (!ev)? group_master_log_pos :
|
2009-12-22 10:35:56 +01:00
|
|
|
((thd->variables.option_bits & OPTION_BEGIN || !ev->log_pos) ?
|
2009-12-14 17:32:22 +01:00
|
|
|
group_master_log_pos : ev->log_pos - ev->data_written);
|
2006-10-31 12:23:14 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{ /* until_condition == UNTIL_RELAY_POS */
|
|
|
|
log_name= group_relay_log_name;
|
|
|
|
log_pos= group_relay_log_pos;
|
|
|
|
}
|
|
|
|
|
2007-01-17 15:06:37 +01:00
|
|
|
#ifndef DBUG_OFF
|
|
|
|
{
|
|
|
|
char buf[32];
|
|
|
|
DBUG_PRINT("info", ("group_master_log_name='%s', group_master_log_pos=%s",
|
|
|
|
group_master_log_name, llstr(group_master_log_pos, buf)));
|
|
|
|
DBUG_PRINT("info", ("group_relay_log_name='%s', group_relay_log_pos=%s",
|
|
|
|
group_relay_log_name, llstr(group_relay_log_pos, buf)));
|
|
|
|
DBUG_PRINT("info", ("(%s) log_name='%s', log_pos=%s",
|
|
|
|
until_condition == UNTIL_MASTER_POS ? "master" : "relay",
|
|
|
|
log_name, llstr(log_pos, buf)));
|
|
|
|
DBUG_PRINT("info", ("(%s) until_log_name='%s', until_log_pos=%s",
|
|
|
|
until_condition == UNTIL_MASTER_POS ? "master" : "relay",
|
|
|
|
until_log_name, llstr(until_log_pos, buf)));
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
if (until_log_names_cmp_result == UNTIL_LOG_NAMES_CMP_UNKNOWN)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
We have no cached comparison results so we should compare log names
|
|
|
|
and cache result.
|
|
|
|
If we are after RESET SLAVE, and the SQL slave thread has not processed
|
|
|
|
any event yet, it could be that group_master_log_name is "". In that case,
|
|
|
|
just wait for more events (as there is no sensible comparison to do).
|
|
|
|
*/
|
|
|
|
|
|
|
|
if (*log_name)
|
|
|
|
{
|
|
|
|
const char *basename= log_name + dirname_length(log_name);
|
|
|
|
|
|
|
|
const char *q= (const char*)(fn_ext(basename)+1);
|
|
|
|
if (strncmp(basename, until_log_name, (int)(q-basename)) == 0)
|
|
|
|
{
|
|
|
|
/* Now compare extensions. */
|
|
|
|
char *q_end;
|
|
|
|
ulong log_name_extension= strtoul(q, &q_end, 10);
|
|
|
|
if (log_name_extension < until_log_name_extension)
|
|
|
|
until_log_names_cmp_result= UNTIL_LOG_NAMES_CMP_LESS;
|
|
|
|
else
|
|
|
|
until_log_names_cmp_result=
|
|
|
|
(log_name_extension > until_log_name_extension) ?
|
|
|
|
UNTIL_LOG_NAMES_CMP_GREATER : UNTIL_LOG_NAMES_CMP_EQUAL ;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
/* Probably error so we aborting */
|
|
|
|
sql_print_error("Slave SQL thread is stopped because UNTIL "
|
|
|
|
"condition is bad.");
|
|
|
|
DBUG_RETURN(TRUE);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
|
|
|
DBUG_RETURN(until_log_pos == 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
DBUG_RETURN(((until_log_names_cmp_result == UNTIL_LOG_NAMES_CMP_EQUAL &&
|
|
|
|
log_pos >= until_log_pos) ||
|
|
|
|
until_log_names_cmp_result == UNTIL_LOG_NAMES_CMP_GREATER));
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
void Relay_log_info::cached_charset_invalidate()
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::cached_charset_invalidate");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
/* Full of zeroes means uninitialized. */
|
|
|
|
bzero(cached_charset, sizeof(cached_charset));
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
bool Relay_log_info::cached_charset_compare(char *charset) const
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::cached_charset_compare");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
Bug#53445: Build with -Wall and fix warnings that it generates
Apart strict-aliasing warnings, fix the remaining warnings
generated by GCC 4.4.4 -Wall and -Wextra flags.
One major source of warnings was the in-house function my_bcmp
which (unconventionally) took pointers to unsigned characters
as the byte sequences to be compared. Since my_bcmp and bcmp
are deprecated functions whose only difference with memcmp is
the return value, every use of the function is replaced with
memcmp as the special return value wasn't actually being used
by any caller.
There were also various other warnings, mostly due to type
mismatches, missing return values, missing prototypes, dead
code (unreachable) and ignored return values.
BUILD/SETUP.sh:
Remove flags that are implied by -Wall and -Wextra.
Do not warn about unused parameters in C++.
BUILD/check-cpu:
Print only the compiler version instead of verbose banner.
Although the option is gcc specific, the check was only
being used for GCC specific checks anyway.
client/mysql.cc:
bcmp is no longer defined.
client/mysqltest.cc:
Pass a string to function expecting a format string.
Replace use of bcmp with memcmp.
cmd-line-utils/readline/Makefile.am:
Always define _GNU_SOURCE when compiling GNU readline.
Required to make certain prototypes visible.
cmd-line-utils/readline/input.c:
Condition for the code to be meaningful.
configure.in:
Remove check for bcmp.
extra/comp_err.c:
Use appropriate type.
extra/replace.c:
Replace use of bcmp with memcmp.
extra/yassl/src/crypto_wrapper.cpp:
Do not ignore the return value of fgets. Retrieve the file
position if fgets succeed -- if it fails, the function will
bail out and return a error.
extra/yassl/taocrypt/include/blowfish.hpp:
Use a single array instead of accessing positions of the sbox_
through a subscript to pbox_.
extra/yassl/taocrypt/include/runtime.hpp:
One definition of such functions is enough.
extra/yassl/taocrypt/src/aes.cpp:
Avoid potentially ambiguous conditions.
extra/yassl/taocrypt/src/algebra.cpp:
Rename arguments to avoid shadowing related warnings.
extra/yassl/taocrypt/src/blowfish.cpp:
Avoid potentially ambiguous conditions.
extra/yassl/taocrypt/src/integer.cpp:
Do not define type within a anonymous union.
Use a variable to return a value instead of
leaving the result in a register -- compiler
does not know the logic inside the asm.
extra/yassl/taocrypt/src/misc.cpp:
Define handler for pure virtual functions.
Remove unused code.
extra/yassl/taocrypt/src/twofish.cpp:
Avoid potentially ambiguous conditions.
extra/yassl/testsuite/test.hpp:
Function must have C language linkage.
include/m_string.h:
Remove check which relied on bcmp being defined -- they weren't
being used as bcmp is only visible when _BSD_SOURCE is defined.
include/my_bitmap.h:
Remove bogus helpers which were used only in a few files and
were causing warnings about dead code.
include/my_global.h:
Due to G++ bug, always silence false-positive uninitialized
variables warnings when compiling C++ code with G++.
Remove bogus helper.
libmysql/Makefile.shared:
Remove built-in implementation of bcmp.
mysql-test/lib/My/SafeProcess/safe_process.cc:
Cast pid to largest possible type for a process identifier.
mysys/mf_loadpath.c:
Leave space of the ending nul.
mysys/mf_pack.c:
Replace bcmp with memcmp.
mysys/my_bitmap.c:
Dead code removal.
mysys/my_gethwaddr.c:
Remove unused variable.
mysys/my_getopt.c:
Silence bogus uninitialized variable warning.
Do not cast away the constant qualifier.
mysys/safemalloc.c:
Cast to expected type.
mysys/thr_lock.c:
Silence bogus uninitialized variable warning.
sql/field.cc:
Replace bogus helper with a more appropriate logic which is
used throughout the code.
sql/item.cc:
Remove bogus logical condition which always evaluates to TRUE.
sql/item_create.cc:
Simplify code to avoid signedness related warnings.
sql/log_event.cc:
Replace use of bcmp with memcmp.
No need to use helpers for simple bit operations.
sql/log_event_old.cc:
Replace bmove_align with memcpy.
sql/mysqld.cc:
Move use declaration of variable to the ifdef block where it
is used. Remove now-unnecessary casts and arguments.
sql/set_var.cc:
Replace bogus helpers with simple and classic bit operations.
sql/slave.cc:
Cast to expected type and silence bogus warning.
sql/sql_class.h:
Don't use enum values as bit flags, the supposed type safety is
bogus as the combined bit flags are not a value in the enumeration.
sql/udf_example.c:
Only declare variable when necessary.
sql/unireg.h:
Replace use of bmove_align with memcpy.
storage/innobase/os/os0file.c:
Silence bogus warning.
storage/myisam/mi_open.c:
Remove bogus cast, DBUG_DUMP expects a pointer to unsigned
char.
storage/myisam/mi_page.c:
Remove bogus cast, DBUG_DUMP expects a pointer to unsigned
char.
strings/bcmp.c:
Remove built-in bcmp.
strings/ctype-ucs2.c:
Silence bogus warning.
tests/mysql_client_test.c:
Use a appropriate type as expected by simple_command().
2010-07-02 20:30:47 +02:00
|
|
|
if (memcmp(cached_charset, charset, sizeof(cached_charset)))
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
2006-11-10 15:10:41 +01:00
|
|
|
memcpy(const_cast<char*>(cached_charset), charset, sizeof(cached_charset));
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
void Relay_log_info::stmt_done(my_off_t event_master_log_pos,
|
2007-04-13 19:19:10 +02:00
|
|
|
time_t event_creation_time)
|
2007-04-12 08:58:04 +02:00
|
|
|
{
|
2007-10-13 22:12:50 +02:00
|
|
|
#ifndef DBUG_OFF
|
|
|
|
extern uint debug_not_change_ts_if_art_event;
|
|
|
|
#endif
|
2007-04-12 08:58:04 +02:00
|
|
|
clear_flag(IN_STMT);
|
|
|
|
|
|
|
|
/*
|
|
|
|
If in a transaction, and if the slave supports transactions, just
|
|
|
|
inc_event_relay_log_pos(). We only have to check for OPTION_BEGIN
|
|
|
|
(not OPTION_NOT_AUTOCOMMIT) as transactions are logged with
|
|
|
|
BEGIN/COMMIT, not with SET AUTOCOMMIT= .
|
|
|
|
|
|
|
|
CAUTION: opt_using_transactions means innodb || bdb ; suppose the
|
|
|
|
master supports InnoDB and BDB, but the slave supports only BDB,
|
|
|
|
problems will arise: - suppose an InnoDB table is created on the
|
|
|
|
master, - then it will be MyISAM on the slave - but as
|
|
|
|
opt_using_transactions is true, the slave will believe he is
|
|
|
|
transactional with the MyISAM table. And problems will come when
|
|
|
|
one does START SLAVE; STOP SLAVE; START SLAVE; (the slave will
|
|
|
|
resume at BEGIN whereas there has not been any rollback). This is
|
|
|
|
the problem of using opt_using_transactions instead of a finer
|
|
|
|
"does the slave support _transactional handler used on the
|
|
|
|
master_".
|
|
|
|
|
|
|
|
More generally, we'll have problems when a query mixes a
|
|
|
|
transactional handler and MyISAM and STOP SLAVE is issued in the
|
|
|
|
middle of the "transaction". START SLAVE will resume at BEGIN
|
|
|
|
while the MyISAM table has already been updated.
|
|
|
|
*/
|
2009-12-22 10:35:56 +01:00
|
|
|
if ((sql_thd->variables.option_bits & OPTION_BEGIN) && opt_using_transactions)
|
2007-04-12 08:58:04 +02:00
|
|
|
inc_event_relay_log_pos();
|
|
|
|
else
|
|
|
|
{
|
|
|
|
inc_group_relay_log_pos(event_master_log_pos);
|
|
|
|
flush_relay_log_info(this);
|
|
|
|
/*
|
|
|
|
Note that Rotate_log_event::do_apply_event() does not call this
|
|
|
|
function, so there is no chance that a fake rotate event resets
|
|
|
|
last_master_timestamp. Note that we update without mutex
|
|
|
|
(probably ok - except in some very rare cases, only consequence
|
|
|
|
is that value may take some time to display in
|
|
|
|
Seconds_Behind_Master - not critical).
|
|
|
|
*/
|
2011-10-19 21:45:18 +02:00
|
|
|
if (!(event_creation_time == 0 &&
|
|
|
|
IF_DBUG(debug_not_change_ts_if_art_event > 0, 1)))
|
2007-10-13 22:12:50 +02:00
|
|
|
last_master_timestamp= event_creation_time;
|
2007-04-12 08:58:04 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
#if !defined(MYSQL_CLIENT) && defined(HAVE_REPLICATION)
|
2007-08-16 07:37:50 +02:00
|
|
|
void Relay_log_info::cleanup_context(THD *thd, bool error)
|
2006-10-31 12:23:14 +01:00
|
|
|
{
|
2007-08-16 07:37:50 +02:00
|
|
|
DBUG_ENTER("Relay_log_info::cleanup_context");
|
2006-10-31 12:23:14 +01:00
|
|
|
|
|
|
|
DBUG_ASSERT(sql_thd == thd);
|
|
|
|
/*
|
2007-03-30 15:29:30 +02:00
|
|
|
1) Instances of Table_map_log_event, if ::do_apply_event() was called on them,
|
2006-10-31 12:23:14 +01:00
|
|
|
may have opened tables, which we cannot be sure have been closed (because
|
|
|
|
maybe the Rows_log_event have not been found or will not be, because slave
|
|
|
|
SQL thread is stopping, or relay log has a missing tail etc). So we close
|
|
|
|
all thread's tables. And so the table mappings have to be cancelled.
|
2007-03-30 15:29:30 +02:00
|
|
|
2) Rows_log_event::do_apply_event() may even have started statements or
|
2006-10-31 12:23:14 +01:00
|
|
|
transactions on them, which we need to rollback in case of error.
|
|
|
|
3) If finding a Format_description_log_event after a BEGIN, we also need
|
|
|
|
to rollback before continuing with the next events.
|
|
|
|
4) so we need this "context cleanup" function.
|
|
|
|
*/
|
|
|
|
if (error)
|
|
|
|
{
|
2009-12-03 19:37:38 +01:00
|
|
|
trans_rollback_stmt(thd); // if a "statement transaction"
|
|
|
|
trans_rollback(thd); // if a "real transaction"
|
2006-10-31 12:23:14 +01:00
|
|
|
}
|
|
|
|
m_table_map.clear_tables();
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
slave_close_thread_tables(thd);
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
if (error)
|
|
|
|
thd->mdl_context.release_transactional_locks();
|
2007-04-12 08:58:04 +02:00
|
|
|
clear_flag(IN_STMT);
|
2007-12-12 11:14:59 +01:00
|
|
|
/*
|
|
|
|
Cleanup for the flags that have been set at do_apply_event.
|
|
|
|
*/
|
2009-12-22 10:35:56 +01:00
|
|
|
thd->variables.option_bits&= ~OPTION_NO_FOREIGN_KEY_CHECKS;
|
|
|
|
thd->variables.option_bits&= ~OPTION_RELAXED_UNIQUE_CHECKS;
|
2011-11-11 18:26:56 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
Reset state related to long_find_row notes in the error log:
|
|
|
|
- timestamp
|
|
|
|
- flag that decides whether the slave prints or not
|
|
|
|
*/
|
|
|
|
reset_row_stmt_start_timestamp();
|
|
|
|
unset_long_find_row_note_printed();
|
|
|
|
|
2006-10-31 12:23:14 +01:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
2007-02-26 17:44:55 +01:00
|
|
|
|
2007-08-16 07:37:50 +02:00
|
|
|
void Relay_log_info::clear_tables_to_lock()
|
2007-02-26 17:44:55 +01:00
|
|
|
{
|
Bug#13693012: SLAVE CRASHING ON INSERT STATEMENT WITH MERGE TABLE
PROBLEM: After WL 4144, when using MyISAM Merge tables, the routine
open_and_lock_tables will append to the list of tables to lock, the
base tables that make up the MERGE table. This has two side-effects in
replication:
1. On the master side, we log additional table maps for the base
tables, since they appear in the list of locked tables, even
though we don't really use them at the slave.
2. On the slave side, when opening a MERGE table while applying a
ROW event, additional tables are appended to the list of tables
to lock.
Side-effect #1 is not harmful. It's just that when using MyISAM Merge
tables a few table maps more may be logged.
Side-effect #2, is harmful, because the list rli->tables_to_lock is an
extended structure from TABLE_LIST in which the extra fields are
filled from the table maps that are processed. Since
open_and_lock_tables appends tables to the list after all table map
events have been processed we end up with entries without
replication/table map data on them. Thus when trying to access that
info for these extra tables, the server will crash.
SOLUTION: We fix side-effect #2 by making sure that we access the
replication part of the structure for those in the list that were
accounted for when processing the correspondent table map events. All
in all, we never go beyond rli->tables_to_lock_count.
We also deploy an assertion when clearing rli->tables_to_lock, making
sure that the base tables are not in the list anymore (were closed in
close_thread_tables).
2012-02-24 17:07:43 +01:00
|
|
|
DBUG_ENTER("Relay_log_info::clear_tables_to_lock()");
|
|
|
|
#ifndef DBUG_OFF
|
|
|
|
/**
|
|
|
|
When replicating in RBR and MyISAM Merge tables are involved
|
|
|
|
open_and_lock_tables (called in do_apply_event) appends the
|
|
|
|
base tables to the list of tables_to_lock. Then these are
|
|
|
|
removed from the list in close_thread_tables (which is called
|
|
|
|
before we reach this point).
|
|
|
|
|
|
|
|
This assertion just confirms that we get no surprises at this
|
|
|
|
point.
|
|
|
|
*/
|
|
|
|
uint i=0;
|
|
|
|
for (TABLE_LIST *ptr= tables_to_lock ; ptr ; ptr= ptr->next_global, i++) ;
|
|
|
|
DBUG_ASSERT(i == tables_to_lock_count);
|
|
|
|
#endif
|
|
|
|
|
2007-02-26 17:44:55 +01:00
|
|
|
while (tables_to_lock)
|
|
|
|
{
|
WL#3817: Simplify string / memory area types and make things more consistent (first part)
The following type conversions was done:
- Changed byte to uchar
- Changed gptr to uchar*
- Change my_string to char *
- Change my_size_t to size_t
- Change size_s to size_t
Removed declaration of byte, gptr, my_string, my_size_t and size_s.
Following function parameter changes was done:
- All string functions in mysys/strings was changed to use size_t
instead of uint for string lengths.
- All read()/write() functions changed to use size_t (including vio).
- All protocoll functions changed to use size_t instead of uint
- Functions that used a pointer to a string length was changed to use size_t*
- Changed malloc(), free() and related functions from using gptr to use void *
as this requires fewer casts in the code and is more in line with how the
standard functions work.
- Added extra length argument to dirname_part() to return the length of the
created string.
- Changed (at least) following functions to take uchar* as argument:
- db_dump()
- my_net_write()
- net_write_command()
- net_store_data()
- DBUG_DUMP()
- decimal2bin() & bin2decimal()
- Changed my_compress() and my_uncompress() to use size_t. Changed one
argument to my_uncompress() from a pointer to a value as we only return
one value (makes function easier to use).
- Changed type of 'pack_data' argument to packfrm() to avoid casts.
- Changed in readfrm() and writefrom(), ha_discover and handler::discover()
the type for argument 'frmdata' to uchar** to avoid casts.
- Changed most Field functions to use uchar* instead of char* (reduced a lot of
casts).
- Changed field->val_xxx(xxx, new_ptr) to take const pointers.
Other changes:
- Removed a lot of not needed casts
- Added a few new cast required by other changes
- Added some cast to my_multi_malloc() arguments for safety (as string lengths
needs to be uint, not size_t).
- Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done
explicitely as this conflict was often hided by casting the function to
hash_get_key).
- Changed some buffers to memory regions to uchar* to avoid casts.
- Changed some string lengths from uint to size_t.
- Changed field->ptr to be uchar* instead of char*. This allowed us to
get rid of a lot of casts.
- Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar
- Include zlib.h in some files as we needed declaration of crc32()
- Changed MY_FILE_ERROR to be (size_t) -1.
- Changed many variables to hold the result of my_read() / my_write() to be
size_t. This was needed to properly detect errors (which are
returned as (size_t) -1).
- Removed some very old VMS code
- Changed packfrm()/unpackfrm() to not be depending on uint size
(portability fix)
- Removed windows specific code to restore cursor position as this
causes slowdown on windows and we should not mix read() and pread()
calls anyway as this is not thread safe. Updated function comment to
reflect this. Changed function that depended on original behavior of
my_pwrite() to itself restore the cursor position (one such case).
- Added some missing checking of return value of malloc().
- Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow.
- Changed type of table_def::m_size from my_size_t to ulong to reflect that
m_size is the number of elements in the array, not a string/memory
length.
- Moved THD::max_row_length() to table.cc (as it's not depending on THD).
Inlined max_row_length_blob() into this function.
- More function comments
- Fixed some compiler warnings when compiled without partitions.
- Removed setting of LEX_STRING() arguments in declaration (portability fix).
- Some trivial indentation/variable name changes.
- Some trivial code simplifications:
- Replaced some calls to alloc_root + memcpy to use
strmake_root()/strdup_root().
- Changed some calls from memdup() to strmake() (Safety fix)
- Simpler loops in client-simple.c
BitKeeper/etc/ignore:
added libmysqld/ha_ndbcluster_cond.cc
---
added debian/defs.mk debian/control
client/completion_hash.cc:
Remove not needed casts
client/my_readline.h:
Remove some old types
client/mysql.cc:
Simplify types
client/mysql_upgrade.c:
Remove some old types
Update call to dirname_part
client/mysqladmin.cc:
Remove some old types
client/mysqlbinlog.cc:
Remove some old types
Change some buffers to be uchar to avoid casts
client/mysqlcheck.c:
Remove some old types
client/mysqldump.c:
Remove some old types
Remove some not needed casts
Change some string lengths to size_t
client/mysqlimport.c:
Remove some old types
client/mysqlshow.c:
Remove some old types
client/mysqlslap.c:
Remove some old types
Remove some not needed casts
client/mysqltest.c:
Removed some old types
Removed some not needed casts
Updated hash-get-key function arguments
Updated parameters to dirname_part()
client/readline.cc:
Removed some old types
Removed some not needed casts
Changed some string lengths to use size_t
client/sql_string.cc:
Removed some old types
dbug/dbug.c:
Removed some old types
Changed some string lengths to use size_t
Changed some prototypes to avoid casts
extra/comp_err.c:
Removed some old types
extra/innochecksum.c:
Removed some old types
extra/my_print_defaults.c:
Removed some old types
extra/mysql_waitpid.c:
Removed some old types
extra/perror.c:
Removed some old types
extra/replace.c:
Removed some old types
Updated parameters to dirname_part()
extra/resolve_stack_dump.c:
Removed some old types
extra/resolveip.c:
Removed some old types
include/config-win.h:
Removed some old types
include/decimal.h:
Changed binary strings to be uchar* instead of char*
include/ft_global.h:
Removed some old types
include/hash.h:
Removed some old types
include/heap.h:
Removed some old types
Changed records_under_level to be 'ulong' instead of 'uint' to clarify usage of variable
include/keycache.h:
Removed some old types
include/m_ctype.h:
Removed some old types
Changed some string lengths to use size_t
Changed character length functions to return uint
unsigned char -> uchar
include/m_string.h:
Removed some old types
Changed some string lengths to use size_t
include/my_alloc.h:
Changed some string lengths to use size_t
include/my_base.h:
Removed some old types
include/my_dbug.h:
Removed some old types
Changed some string lengths to use size_t
Changed db_dump() to take uchar * as argument for memory to reduce number of casts in usage
include/my_getopt.h:
Removed some old types
include/my_global.h:
Removed old types:
my_size_t -> size_t
byte -> uchar
gptr -> uchar *
include/my_list.h:
Removed some old types
include/my_nosys.h:
Removed some old types
include/my_pthread.h:
Removed some old types
include/my_sys.h:
Removed some old types
Changed MY_FILE_ERROR to be in line with new definitions of my_write()/my_read()
Changed some string lengths to use size_t
my_malloc() / my_free() now uses void *
Updated parameters to dirname_part() & my_uncompress()
include/my_tree.h:
Removed some old types
include/my_trie.h:
Removed some old types
include/my_user.h:
Changed some string lengths to use size_t
include/my_vle.h:
Removed some old types
include/my_xml.h:
Removed some old types
Changed some string lengths to use size_t
include/myisam.h:
Removed some old types
include/myisammrg.h:
Removed some old types
include/mysql.h:
Removed some old types
Changed byte streams to use uchar* instead of char*
include/mysql_com.h:
Removed some old types
Changed some string lengths to use size_t
Changed some buffers to be uchar* to avoid casts
include/queues.h:
Removed some old types
include/sql_common.h:
Removed some old types
include/sslopt-longopts.h:
Removed some old types
include/violite.h:
Removed some old types
Changed some string lengths to use size_t
libmysql/client_settings.h:
Removed some old types
libmysql/libmysql.c:
Removed some old types
libmysql/manager.c:
Removed some old types
libmysqld/emb_qcache.cc:
Removed some old types
libmysqld/emb_qcache.h:
Removed some old types
libmysqld/lib_sql.cc:
Removed some old types
Removed some not needed casts
Changed some buffers to be uchar* to avoid casts
true -> TRUE, false -> FALSE
mysys/array.c:
Removed some old types
mysys/charset.c:
Changed some string lengths to use size_t
mysys/checksum.c:
Include zlib to get definition for crc32
Removed some old types
mysys/default.c:
Removed some old types
Changed some string lengths to use size_t
mysys/default_modify.c:
Changed some string lengths to use size_t
Removed some not needed casts
mysys/hash.c:
Removed some old types
Changed some string lengths to use size_t
Note: Prototype of hash_key() has changed which may cause problems if client uses hash_init() with a cast for the hash-get-key function.
hash_element now takes 'ulong' as the index type (cleanup)
mysys/list.c:
Removed some old types
mysys/mf_cache.c:
Changed some string lengths to use size_t
mysys/mf_dirname.c:
Removed some old types
Changed some string lengths to use size_t
Added argument to dirname_part() to avoid calculation of length for 'to'
mysys/mf_fn_ext.c:
Removed some old types
Updated parameters to dirname_part()
mysys/mf_format.c:
Removed some old types
Changed some string lengths to use size_t
mysys/mf_getdate.c:
Removed some old types
mysys/mf_iocache.c:
Removed some old types
Changed some string lengths to use size_t
Changed calculation of 'max_length' to be done the same way in all functions
mysys/mf_iocache2.c:
Removed some old types
Changed some string lengths to use size_t
Clean up comments
Removed not needed indentation
mysys/mf_keycache.c:
Removed some old types
mysys/mf_keycaches.c:
Removed some old types
mysys/mf_loadpath.c:
Removed some old types
mysys/mf_pack.c:
Removed some old types
Changed some string lengths to use size_t
Removed some not needed casts
Removed very old VMS code
Updated parameters to dirname_part()
Use result of dirnam_part() to remove call to strcat()
mysys/mf_path.c:
Removed some old types
mysys/mf_radix.c:
Removed some old types
mysys/mf_same.c:
Removed some old types
mysys/mf_sort.c:
Removed some old types
mysys/mf_soundex.c:
Removed some old types
mysys/mf_strip.c:
Removed some old types
mysys/mf_tempdir.c:
Removed some old types
mysys/mf_unixpath.c:
Removed some old types
mysys/mf_wfile.c:
Removed some old types
mysys/mulalloc.c:
Removed some old types
mysys/my_alloc.c:
Removed some old types
Changed some string lengths to use size_t
Use void* as type for allocated memory area
Removed some not needed casts
Changed argument 'Size' to 'length' according coding guidelines
mysys/my_chsize.c:
Changed some buffers to be uchar* to avoid casts
mysys/my_compress.c:
More comments
Removed some old types
Changed string lengths to use size_t
Changed arguments to my_uncompress() to make them easier to understand
Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix)
Changed type of 'pack_data' argument to packfrm() to avoid casts.
mysys/my_conio.c:
Changed some string lengths to use size_t
mysys/my_create.c:
Removed some old types
mysys/my_div.c:
Removed some old types
mysys/my_error.c:
Removed some old types
mysys/my_fopen.c:
Removed some old types
mysys/my_fstream.c:
Removed some old types
Changed some string lengths to use size_t
writen -> written
mysys/my_getopt.c:
Removed some old types
mysys/my_getwd.c:
Removed some old types
More comments
mysys/my_init.c:
Removed some old types
mysys/my_largepage.c:
Removed some old types
Changed some string lengths to use size_t
mysys/my_lib.c:
Removed some old types
mysys/my_lockmem.c:
Removed some old types
mysys/my_malloc.c:
Removed some old types
Changed malloc(), free() and related functions to use void *
Changed all functions to use size_t
mysys/my_memmem.c:
Indentation cleanup
mysys/my_once.c:
Removed some old types
Changed malloc(), free() and related functions to use void *
mysys/my_open.c:
Removed some old types
mysys/my_pread.c:
Removed some old types
Changed all functions to use size_t
Added comment for how my_pread() / my_pwrite() are supposed to work.
Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe.
(If we ever would really need this, it should be enabled only with a flag argument)
mysys/my_quick.c:
Removed some old types
Changed all functions to use size_t
mysys/my_read.c:
Removed some old types
Changed all functions to use size_t
mysys/my_realloc.c:
Removed some old types
Use void* as type for allocated memory area
Changed all functions to use size_t
mysys/my_static.c:
Removed some old types
mysys/my_static.h:
Removed some old types
mysys/my_vle.c:
Removed some old types
mysys/my_wincond.c:
Removed some old types
mysys/my_windac.c:
Removed some old types
mysys/my_write.c:
Removed some old types
Changed all functions to use size_t
mysys/ptr_cmp.c:
Removed some old types
Changed all functions to use size_t
mysys/queues.c:
Removed some old types
mysys/safemalloc.c:
Removed some old types
Changed malloc(), free() and related functions to use void *
Changed all functions to use size_t
mysys/string.c:
Removed some old types
Changed all functions to use size_t
mysys/testhash.c:
Removed some old types
mysys/thr_alarm.c:
Removed some old types
mysys/thr_lock.c:
Removed some old types
mysys/tree.c:
Removed some old types
mysys/trie.c:
Removed some old types
mysys/typelib.c:
Removed some old types
plugin/daemon_example/daemon_example.cc:
Removed some old types
regex/reginit.c:
Removed some old types
server-tools/instance-manager/buffer.cc:
Changed some string lengths to use size_t
Changed buffer to be of type uchar*
server-tools/instance-manager/buffer.h:
Changed some string lengths to use size_t
Changed buffer to be of type uchar*
server-tools/instance-manager/commands.cc:
Removed some old types
Changed some string lengths to use size_t
Changed buffer to be of type uchar*
server-tools/instance-manager/instance_map.cc:
Removed some old types
Changed some string lengths to use size_t
Changed buffer to be of type uchar*
server-tools/instance-manager/instance_options.cc:
Changed buffer to be of type uchar*
Replaced alloc_root + strcpy() with strdup_root()
server-tools/instance-manager/mysql_connection.cc:
Changed buffer to be of type uchar*
server-tools/instance-manager/options.cc:
Removed some old types
server-tools/instance-manager/parse.cc:
Changed some string lengths to use size_t
server-tools/instance-manager/parse.h:
Removed some old types
Changed some string lengths to use size_t
server-tools/instance-manager/protocol.cc:
Changed some buffers to be uchar* to avoid casts
Changed some string lengths to use size_t
server-tools/instance-manager/protocol.h:
Changed some string lengths to use size_t
server-tools/instance-manager/user_map.cc:
Removed some old types
Changed some string lengths to use size_t
sql/derror.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Changed some string lengths to use size_t
sql/discover.cc:
Changed in readfrm() and writefrom() the type for argument 'frmdata' to uchar** to avoid casts
Changed some string lengths to use size_t
Changed some buffers to be uchar* to avoid casts
sql/event_data_objects.cc:
Removed some old types
Added missing casts for alloc() and sprintf()
sql/event_db_repository.cc:
Changed some buffers to be uchar* to avoid casts
Added missing casts for sprintf()
sql/event_queue.cc:
Removed some old types
sql/field.cc:
Removed some old types
Changed memory buffers to be uchar*
Changed some string lengths to use size_t
Removed a lot of casts
Safety fix in Field_blob::val_decimal() to not access zero pointer
sql/field.h:
Removed some old types
Changed memory buffers to be uchar* (except of store() as this would have caused too many other changes).
Changed some string lengths to use size_t
Removed some not needed casts
Changed val_xxx(xxx, new_ptr) to take const pointers
sql/field_conv.cc:
Removed some old types
Added casts required because memory area pointers are now uchar*
sql/filesort.cc:
Initalize variable that was used unitialized in error conditions
sql/gen_lex_hash.cc:
Removed some old types
Changed memory buffers to be uchar*
Changed some string lengths to use size_t
Removed a lot of casts
Safety fix in Field_blob::val_decimal() to not access zero pointer
sql/gstream.h:
Added required cast
sql/ha_ndbcluster.cc:
Removed some old types
Updated hash-get-key function arguments
Changed some buffers to be uchar* to avoid casts
Added required casts
Removed some not needed casts
sql/ha_ndbcluster.h:
Removed some old types
sql/ha_ndbcluster_binlog.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Replaced sql_alloc() + memcpy() + set end 0 with sql_strmake()
Changed some string lengths to use size_t
Added missing casts for alloc() and sprintf()
sql/ha_ndbcluster_binlog.h:
Removed some old types
sql/ha_ndbcluster_cond.cc:
Removed some old types
Removed some not needed casts
sql/ha_ndbcluster_cond.h:
Removed some old types
sql/ha_partition.cc:
Removed some old types
Changed prototype for change_partition() to avoid casts
sql/ha_partition.h:
Removed some old types
sql/handler.cc:
Removed some old types
Changed some string lengths to use size_t
sql/handler.h:
Removed some old types
Changed some string lengths to use size_t
Changed type for 'frmblob' parameter for discover() and ha_discover() to get fewer casts
sql/hash_filo.h:
Removed some old types
Changed all functions to use size_t
sql/hostname.cc:
Removed some old types
sql/item.cc:
Removed some old types
Changed some string lengths to use size_t
Use strmake() instead of memdup() to create a null terminated string.
Updated calls to new Field()
sql/item.h:
Removed some old types
Changed malloc(), free() and related functions to use void *
Changed some buffers to be uchar* to avoid casts
sql/item_cmpfunc.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
sql/item_cmpfunc.h:
Removed some old types
sql/item_create.cc:
Removed some old types
sql/item_func.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
Added test for failing alloc() in init_result_field()
Remove old confusing comment
Fixed compiler warning
sql/item_func.h:
Removed some old types
sql/item_row.cc:
Removed some old types
sql/item_row.h:
Removed some old types
sql/item_strfunc.cc:
Include zlib (needed becasue we call crc32)
Removed some old types
sql/item_strfunc.h:
Removed some old types
Changed some types to match new function prototypes
sql/item_subselect.cc:
Removed some old types
sql/item_subselect.h:
Removed some old types
sql/item_sum.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
sql/item_sum.h:
Removed some old types
sql/item_timefunc.cc:
Removed some old types
Changed some string lengths to use size_t
sql/item_timefunc.h:
Removed some old types
sql/item_xmlfunc.cc:
Changed some string lengths to use size_t
sql/item_xmlfunc.h:
Removed some old types
sql/key.cc:
Removed some old types
Removed some not needed casts
sql/lock.cc:
Removed some old types
Added some cast to my_multi_malloc() arguments for safety
sql/log.cc:
Removed some old types
Changed some string lengths to use size_t
Changed some buffers to be uchar* to avoid casts
Changed usage of pwrite() to not assume it holds the cursor position for the file
Made usage of my_read() safer
sql/log_event.cc:
Removed some old types
Added checking of return value of malloc() in pack_info()
Changed some buffers to be uchar* to avoid casts
Removed some 'const' to avoid casts
Added missing casts for alloc() and sprintf()
Added required casts
Removed some not needed casts
Added some cast to my_multi_malloc() arguments for safety
sql/log_event.h:
Removed some old types
Changed some buffers to be uchar* to avoid casts
sql/log_event_old.cc:
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
sql/log_event_old.h:
Changed some buffers to be uchar* to avoid casts
sql/mf_iocache.cc:
Removed some old types
sql/my_decimal.cc:
Changed memory area to use uchar*
sql/my_decimal.h:
Changed memory area to use uchar*
sql/mysql_priv.h:
Removed some old types
Changed malloc(), free() and related functions to use void *
Changed some string lengths to use size_t
Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid long overflow
Changed some buffers to be uchar* to avoid casts
sql/mysqld.cc:
Removed some old types
sql/net_serv.cc:
Removed some old types
Changed some string lengths to use size_t
Changed some buffers to be uchar* to avoid casts
Ensure that vio_read()/vio_write() return values are stored in a size_t variable
Removed some not needed casts
sql/opt_range.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
sql/opt_range.h:
Removed some old types
Changed some buffers to be uchar* to avoid casts
sql/opt_sum.cc:
Removed some old types
Removed some not needed casts
sql/parse_file.cc:
Removed some old types
Changed some string lengths to use size_t
Changed alloc_root + memcpy + set end 0 -> strmake_root()
sql/parse_file.h:
Removed some old types
sql/partition_info.cc:
Removed some old types
Added missing casts for alloc()
Changed some buffers to be uchar* to avoid casts
sql/partition_info.h:
Changed some buffers to be uchar* to avoid casts
sql/protocol.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
sql/protocol.h:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Changed some string lengths to use size_t
sql/records.cc:
Removed some old types
sql/repl_failsafe.cc:
Removed some old types
Changed some string lengths to use size_t
Added required casts
sql/rpl_filter.cc:
Removed some old types
Updated hash-get-key function arguments
Changed some string lengths to use size_t
sql/rpl_filter.h:
Changed some string lengths to use size_t
sql/rpl_injector.h:
Removed some old types
sql/rpl_record.cc:
Removed some old types
Removed some not needed casts
Changed some buffers to be uchar* to avoid casts
sql/rpl_record.h:
Removed some old types
Changed some buffers to be uchar* to avoid casts
sql/rpl_record_old.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
sql/rpl_record_old.h:
Removed some old types
Changed some buffers to be uchar* to avoid cast
sql/rpl_rli.cc:
Removed some old types
sql/rpl_tblmap.cc:
Removed some old types
sql/rpl_tblmap.h:
Removed some old types
sql/rpl_utility.cc:
Removed some old types
sql/rpl_utility.h:
Removed some old types
Changed type of m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length
sql/set_var.cc:
Removed some old types
Updated parameters to dirname_part()
sql/set_var.h:
Removed some old types
sql/slave.cc:
Removed some old types
Changed some string lengths to use size_t
sql/slave.h:
Removed some old types
sql/sp.cc:
Removed some old types
Added missing casts for printf()
sql/sp.h:
Removed some old types
Updated hash-get-key function arguments
sql/sp_cache.cc:
Removed some old types
Added missing casts for printf()
Updated hash-get-key function arguments
sql/sp_head.cc:
Removed some old types
Added missing casts for alloc() and printf()
Added required casts
Updated hash-get-key function arguments
sql/sp_head.h:
Removed some old types
sql/sp_pcontext.cc:
Removed some old types
sql/sp_pcontext.h:
Removed some old types
sql/sql_acl.cc:
Removed some old types
Changed some string lengths to use size_t
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
Added required casts
sql/sql_analyse.cc:
Changed some buffers to be uchar* to avoid casts
sql/sql_analyse.h:
Changed some buffers to be uchar* to avoid casts
sql/sql_array.h:
Removed some old types
sql/sql_base.cc:
Removed some old types
Updated hash-get-key function arguments
sql/sql_binlog.cc:
Removed some old types
Added missing casts for printf()
sql/sql_cache.cc:
Removed some old types
Updated hash-get-key function arguments
Removed some not needed casts
Changed some string lengths to use size_t
sql/sql_cache.h:
Removed some old types
Removed reference to not existing function cache_key()
Updated hash-get-key function arguments
sql/sql_class.cc:
Removed some old types
Updated hash-get-key function arguments
Added missing casts for alloc()
Updated hash-get-key function arguments
Moved THD::max_row_length() to table.cc (as it's not depending on THD)
Removed some not needed casts
sql/sql_class.h:
Removed some old types
Changed malloc(), free() and related functions to use void *
Removed some not needed casts
Changed some string lengths to use size_t
Moved max_row_length and max_row_length_blob() to table.cc, as they are not depending on THD
sql/sql_connect.cc:
Removed some old types
Added required casts
sql/sql_db.cc:
Removed some old types
Removed some not needed casts
Added some cast to my_multi_malloc() arguments for safety
Added missing casts for alloc()
sql/sql_delete.cc:
Removed some old types
sql/sql_handler.cc:
Removed some old types
Updated hash-get-key function arguments
Added some cast to my_multi_malloc() arguments for safety
sql/sql_help.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
sql/sql_insert.cc:
Removed some old types
Added missing casts for alloc() and printf()
sql/sql_lex.cc:
Removed some old types
sql/sql_lex.h:
Removed some old types
Removed some not needed casts
sql/sql_list.h:
Removed some old types
Removed some not needed casts
sql/sql_load.cc:
Removed some old types
Removed compiler warning
sql/sql_manager.cc:
Removed some old types
sql/sql_map.cc:
Removed some old types
sql/sql_map.h:
Removed some old types
sql/sql_olap.cc:
Removed some old types
sql/sql_parse.cc:
Removed some old types
Trivial move of code lines to make things more readable
Changed some string lengths to use size_t
Added missing casts for alloc()
sql/sql_partition.cc:
Removed some old types
Removed compiler warnings about not used functions
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
sql/sql_partition.h:
Removed some old types
Changed some buffers to be uchar* to avoid casts
sql/sql_plugin.cc:
Removed some old types
Added missing casts for alloc()
Updated hash-get-key function arguments
sql/sql_prepare.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Added missing casts for alloc() and printf()
sql-common/client.c:
Removed some old types
Changed some memory areas to use uchar*
sql-common/my_user.c:
Changed some string lengths to use size_t
sql-common/pack.c:
Changed some buffers to be uchar* to avoid casts
sql/sql_repl.cc:
Added required casts
Changed some buffers to be uchar* to avoid casts
Changed some string lengths to use size_t
sql/sql_select.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Removed some old types
sql/sql_select.h:
Removed some old types
Changed some buffers to be uchar* to avoid casts
sql/sql_servers.cc:
Removed some old types
Updated hash-get-key function arguments
sql/sql_show.cc:
Removed some old types
Added missing casts for alloc()
Removed some not needed casts
sql/sql_string.cc:
Removed some old types
Added required casts
sql/sql_table.cc:
Removed some old types
Removed compiler warning about not used variable
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
sql/sql_test.cc:
Removed some old types
sql/sql_trigger.cc:
Removed some old types
Added missing casts for alloc()
sql/sql_udf.cc:
Removed some old types
Updated hash-get-key function arguments
sql/sql_union.cc:
Removed some old types
sql/sql_update.cc:
Removed some old types
Removed some not needed casts
sql/sql_view.cc:
Removed some old types
sql/sql_yacc.yy:
Removed some old types
Changed some string lengths to use size_t
Added missing casts for alloc()
sql/stacktrace.c:
Removed some old types
sql/stacktrace.h:
Removed some old types
sql/structs.h:
Removed some old types
sql/table.cc:
Removed some old types
Updated hash-get-key function arguments
Changed some buffers to be uchar* to avoid casts
Removed setting of LEX_STRING() arguments in declaration
Added required casts
More function comments
Moved max_row_length() here from sql_class.cc/sql_class.h
sql/table.h:
Removed some old types
Changed some string lengths to use size_t
sql/thr_malloc.cc:
Use void* as type for allocated memory area
Changed all functions to use size_t
sql/tzfile.h:
Changed some buffers to be uchar* to avoid casts
sql/tztime.cc:
Changed some buffers to be uchar* to avoid casts
Updated hash-get-key function arguments
Added missing casts for alloc()
Removed some not needed casts
sql/uniques.cc:
Removed some old types
Removed some not needed casts
sql/unireg.cc:
Removed some old types
Changed some buffers to be uchar* to avoid casts
Removed some not needed casts
Added missing casts for alloc()
storage/archive/archive_reader.c:
Removed some old types
storage/archive/azio.c:
Removed some old types
Removed some not needed casts
storage/archive/ha_archive.cc:
Removed some old types
Changed type for 'frmblob' in archive_discover() to match handler
Updated hash-get-key function arguments
Removed some not needed casts
storage/archive/ha_archive.h:
Removed some old types
storage/blackhole/ha_blackhole.cc:
Removed some old types
storage/blackhole/ha_blackhole.h:
Removed some old types
storage/csv/ha_tina.cc:
Removed some old types
Updated hash-get-key function arguments
Changed some buffers to be uchar* to avoid casts
storage/csv/ha_tina.h:
Removed some old types
Removed some not needed casts
storage/csv/transparent_file.cc:
Removed some old types
Changed type of 'bytes_read' to be able to detect read errors
Fixed indentation
storage/csv/transparent_file.h:
Removed some old types
storage/example/ha_example.cc:
Removed some old types
Updated hash-get-key function arguments
storage/example/ha_example.h:
Removed some old types
storage/federated/ha_federated.cc:
Removed some old types
Updated hash-get-key function arguments
Removed some not needed casts
storage/federated/ha_federated.h:
Removed some old types
storage/heap/_check.c:
Changed some buffers to be uchar* to avoid casts
storage/heap/_rectest.c:
Removed some old types
storage/heap/ha_heap.cc:
Removed some old types
storage/heap/ha_heap.h:
Removed some old types
storage/heap/heapdef.h:
Removed some old types
storage/heap/hp_block.c:
Removed some old types
Changed some string lengths to use size_t
storage/heap/hp_clear.c:
Removed some old types
storage/heap/hp_close.c:
Removed some old types
storage/heap/hp_create.c:
Removed some old types
storage/heap/hp_delete.c:
Removed some old types
storage/heap/hp_hash.c:
Removed some old types
storage/heap/hp_info.c:
Removed some old types
storage/heap/hp_open.c:
Removed some old types
storage/heap/hp_rfirst.c:
Removed some old types
storage/heap/hp_rkey.c:
Removed some old types
storage/heap/hp_rlast.c:
Removed some old types
storage/heap/hp_rnext.c:
Removed some old types
storage/heap/hp_rprev.c:
Removed some old types
storage/heap/hp_rrnd.c:
Removed some old types
storage/heap/hp_rsame.c:
Removed some old types
storage/heap/hp_scan.c:
Removed some old types
storage/heap/hp_test1.c:
Removed some old types
storage/heap/hp_test2.c:
Removed some old types
storage/heap/hp_update.c:
Removed some old types
storage/heap/hp_write.c:
Removed some old types
Changed some string lengths to use size_t
storage/innobase/handler/ha_innodb.cc:
Removed some old types
Updated hash-get-key function arguments
Added missing casts for alloc() and printf()
Removed some not needed casts
storage/innobase/handler/ha_innodb.h:
Removed some old types
storage/myisam/ft_boolean_search.c:
Removed some old types
storage/myisam/ft_nlq_search.c:
Removed some old types
storage/myisam/ft_parser.c:
Removed some old types
Changed some buffers to be uchar* to avoid casts
storage/myisam/ft_static.c:
Removed some old types
storage/myisam/ft_stopwords.c:
Removed some old types
storage/myisam/ft_update.c:
Removed some old types
Changed some buffers to be uchar* to avoid casts
storage/myisam/ftdefs.h:
Removed some old types
Changed some buffers to be uchar* to avoid casts
storage/myisam/fulltext.h:
Removed some old types
storage/myisam/ha_myisam.cc:
Removed some old types
storage/myisam/ha_myisam.h:
Removed some old types
storage/myisam/mi_cache.c:
Removed some old types
Changed some buffers to be uchar* to avoid casts
storage/myisam/mi_check.c:
Removed some old types
storage/myisam/mi_checksum.c:
Removed some old types
storage/myisam/mi_close.c:
Removed some old types
storage/myisam/mi_create.c:
Removed some old types
storage/myisam/mi_delete.c:
Removed some old types
storage/myisam/mi_delete_all.c:
Removed some old types
storage/myisam/mi_dynrec.c:
Removed some old types
storage/myisam/mi_extra.c:
Removed some old types
storage/myisam/mi_key.c:
Removed some old types
storage/myisam/mi_locking.c:
Removed some old types
storage/myisam/mi_log.c:
Removed some old types
storage/myisam/mi_open.c:
Removed some old types
Removed some not needed casts
Check argument of my_write()/my_pwrite() in functions returning int
Added casting of string lengths to size_t
storage/myisam/mi_packrec.c:
Removed some old types
Changed some buffers to be uchar* to avoid casts
storage/myisam/mi_page.c:
Removed some old types
storage/myisam/mi_preload.c:
Removed some old types
storage/myisam/mi_range.c:
Removed some old types
storage/myisam/mi_rfirst.c:
Removed some old types
storage/myisam/mi_rkey.c:
Removed some old types
storage/myisam/mi_rlast.c:
Removed some old types
storage/myisam/mi_rnext.c:
Removed some old types
storage/myisam/mi_rnext_same.c:
Removed some old types
storage/myisam/mi_rprev.c:
Removed some old types
storage/myisam/mi_rrnd.c:
Removed some old types
storage/myisam/mi_rsame.c:
Removed some old types
storage/myisam/mi_rsamepos.c:
Removed some old types
storage/myisam/mi_scan.c:
Removed some old types
storage/myisam/mi_search.c:
Removed some old types
storage/myisam/mi_static.c:
Removed some old types
storage/myisam/mi_statrec.c:
Removed some old types
storage/myisam/mi_test1.c:
Removed some old types
storage/myisam/mi_test2.c:
Removed some old types
storage/myisam/mi_test3.c:
Removed some old types
storage/myisam/mi_unique.c:
Removed some old types
storage/myisam/mi_update.c:
Removed some old types
storage/myisam/mi_write.c:
Removed some old types
storage/myisam/myisam_ftdump.c:
Removed some old types
storage/myisam/myisamchk.c:
Removed some old types
storage/myisam/myisamdef.h:
Removed some old types
storage/myisam/myisamlog.c:
Removed some old types
Indentation fix
storage/myisam/myisampack.c:
Removed some old types
storage/myisam/rt_index.c:
Removed some old types
storage/myisam/rt_split.c:
Removed some old types
storage/myisam/sort.c:
Removed some old types
storage/myisam/sp_defs.h:
Removed some old types
storage/myisam/sp_key.c:
Removed some old types
storage/myisammrg/ha_myisammrg.cc:
Removed some old types
storage/myisammrg/ha_myisammrg.h:
Removed some old types
storage/myisammrg/myrg_close.c:
Removed some old types
storage/myisammrg/myrg_def.h:
Removed some old types
storage/myisammrg/myrg_delete.c:
Removed some old types
storage/myisammrg/myrg_open.c:
Removed some old types
Updated parameters to dirname_part()
storage/myisammrg/myrg_queue.c:
Removed some old types
storage/myisammrg/myrg_rfirst.c:
Removed some old types
storage/myisammrg/myrg_rkey.c:
Removed some old types
storage/myisammrg/myrg_rlast.c:
Removed some old types
storage/myisammrg/myrg_rnext.c:
Removed some old types
storage/myisammrg/myrg_rnext_same.c:
Removed some old types
storage/myisammrg/myrg_rprev.c:
Removed some old types
storage/myisammrg/myrg_rrnd.c:
Removed some old types
storage/myisammrg/myrg_rsame.c:
Removed some old types
storage/myisammrg/myrg_update.c:
Removed some old types
storage/myisammrg/myrg_write.c:
Removed some old types
storage/ndb/include/util/ndb_opts.h:
Removed some old types
storage/ndb/src/cw/cpcd/main.cpp:
Removed some old types
storage/ndb/src/kernel/vm/Configuration.cpp:
Removed some old types
storage/ndb/src/mgmclient/main.cpp:
Removed some old types
storage/ndb/src/mgmsrv/InitConfigFileParser.cpp:
Removed some old types
Removed old disabled code
storage/ndb/src/mgmsrv/main.cpp:
Removed some old types
storage/ndb/src/ndbapi/NdbBlob.cpp:
Removed some old types
storage/ndb/src/ndbapi/NdbOperationDefine.cpp:
Removed not used variable
storage/ndb/src/ndbapi/NdbOperationInt.cpp:
Added required casts
storage/ndb/src/ndbapi/NdbScanOperation.cpp:
Added required casts
storage/ndb/tools/delete_all.cpp:
Removed some old types
storage/ndb/tools/desc.cpp:
Removed some old types
storage/ndb/tools/drop_index.cpp:
Removed some old types
storage/ndb/tools/drop_tab.cpp:
Removed some old types
storage/ndb/tools/listTables.cpp:
Removed some old types
storage/ndb/tools/ndb_config.cpp:
Removed some old types
storage/ndb/tools/restore/consumer_restore.cpp:
Changed some buffers to be uchar* to avoid casts with new defintion of packfrm()
storage/ndb/tools/restore/restore_main.cpp:
Removed some old types
storage/ndb/tools/select_all.cpp:
Removed some old types
storage/ndb/tools/select_count.cpp:
Removed some old types
storage/ndb/tools/waiter.cpp:
Removed some old types
strings/bchange.c:
Changed function to use uchar * and size_t
strings/bcmp.c:
Changed function to use uchar * and size_t
strings/bmove512.c:
Changed function to use uchar * and size_t
strings/bmove_upp.c:
Changed function to use uchar * and size_t
strings/ctype-big5.c:
Changed functions to use size_t
Changed character length functions to return uint
strings/ctype-bin.c:
Changed functions to use size_t
strings/ctype-cp932.c:
Changed functions to use size_t
Changed character length functions to return uint
strings/ctype-czech.c:
Fixed indentation
Changed functions to use size_t
strings/ctype-euc_kr.c:
Changed functions to use size_t
Changed character length functions to return uint
strings/ctype-eucjpms.c:
Changed functions to use size_t
Changed character length functions to return uint
unsigned char -> uchar
strings/ctype-gb2312.c:
Changed functions to use size_t
Changed character length functions to return uint
strings/ctype-gbk.c:
Changed functions to use size_t
Changed character length functions to return uint
strings/ctype-latin1.c:
Changed functions to use size_t
Changed character length functions to return uint
unsigned char -> uchar
strings/ctype-mb.c:
Changed functions to use size_t
Changed character length functions to return uint
strings/ctype-simple.c:
Changed functions to use size_t
Simpler loops for caseup/casedown
unsigned int -> uint
unsigned char -> uchar
strings/ctype-sjis.c:
Changed functions to use size_t
Changed character length functions to return uint
strings/ctype-tis620.c:
Changed functions to use size_t
Changed character length functions to return uint
unsigned char -> uchar
strings/ctype-uca.c:
Changed functions to use size_t
unsigned char -> uchar
strings/ctype-ucs2.c:
Moved inclusion of stdarg.h to other includes
usigned char -> uchar
Changed functions to use size_t
Changed character length functions to return uint
strings/ctype-ujis.c:
Changed functions to use size_t
Changed character length functions to return uint
unsigned char -> uchar
strings/ctype-utf8.c:
Changed functions to use size_t
unsigned char -> uchar
Indentation fixes
strings/ctype-win1250ch.c:
Indentation fixes
Changed functions to use size_t
strings/ctype.c:
Changed functions to use size_t
strings/decimal.c:
Changed type for memory argument to uchar *
strings/do_ctype.c:
Indentation fixes
strings/my_strtoll10.c:
unsigned char -> uchar
strings/my_vsnprintf.c:
Changed functions to use size_t
strings/r_strinstr.c:
Removed some old types
Changed functions to use size_t
strings/str_test.c:
Removed some old types
strings/strappend.c:
Changed functions to use size_t
strings/strcont.c:
Removed some old types
strings/strfill.c:
Removed some old types
strings/strinstr.c:
Changed functions to use size_t
strings/strlen.c:
Changed functions to use size_t
strings/strmake.c:
Changed functions to use size_t
strings/strnlen.c:
Changed functions to use size_t
strings/strnmov.c:
Changed functions to use size_t
strings/strto.c:
unsigned char -> uchar
strings/strtod.c:
Changed functions to use size_t
strings/strxnmov.c:
Changed functions to use size_t
strings/xml.c:
Changed functions to use size_t
Indentation fixes
tests/mysql_client_test.c:
Removed some old types
tests/thread_test.c:
Removed some old types
vio/test-ssl.c:
Removed some old types
vio/test-sslclient.c:
Removed some old types
vio/test-sslserver.c:
Removed some old types
vio/vio.c:
Removed some old types
vio/vio_priv.h:
Removed some old types
Changed vio_read()/vio_write() to work with size_t
vio/viosocket.c:
Changed vio_read()/vio_write() to work with size_t
Indentation fixes
vio/viossl.c:
Changed vio_read()/vio_write() to work with size_t
Indentation fixes
vio/viosslfactories.c:
Removed some old types
vio/viotest-ssl.c:
Removed some old types
win/README:
More explanations
2007-05-10 11:59:39 +02:00
|
|
|
uchar* to_free= reinterpret_cast<uchar*>(tables_to_lock);
|
2007-02-26 17:44:55 +01:00
|
|
|
if (tables_to_lock->m_tabledef_valid)
|
|
|
|
{
|
|
|
|
tables_to_lock->m_tabledef.table_def::~table_def();
|
|
|
|
tables_to_lock->m_tabledef_valid= FALSE;
|
|
|
|
}
|
2011-05-24 00:46:51 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
If blob fields were used during conversion of field values
|
|
|
|
from the master table into the slave table, then we need to
|
|
|
|
free the memory used temporarily to store their values before
|
|
|
|
copying into the slave's table.
|
|
|
|
*/
|
|
|
|
if (tables_to_lock->m_conv_table)
|
|
|
|
free_blobs(tables_to_lock->m_conv_table);
|
|
|
|
|
2007-02-26 17:44:55 +01:00
|
|
|
tables_to_lock=
|
|
|
|
static_cast<RPL_TABLE_LIST*>(tables_to_lock->next_global);
|
|
|
|
tables_to_lock_count--;
|
Bug#34043: Server loops excessively in _checkchunk() when safemalloc is enabled
Essentially, the problem is that safemalloc is excruciatingly
slow as it checks all allocated blocks for overrun at each
memory management primitive, yielding a almost exponential
slowdown for the memory management functions (malloc, realloc,
free). The overrun check basically consists of verifying some
bytes of a block for certain magic keys, which catches some
simple forms of overrun. Another minor problem is violation
of aliasing rules and that its own internal list of blocks
is prone to corruption.
Another issue with safemalloc is rather the maintenance cost
as the tool has a significant impact on the server code.
Given the magnitude of memory debuggers available nowadays,
especially those that are provided with the platform malloc
implementation, maintenance of a in-house and largely obsolete
memory debugger becomes a burden that is not worth the effort
due to its slowness and lack of support for detecting more
common forms of heap corruption.
Since there are third-party tools that can provide the same
functionality at a lower or comparable performance cost, the
solution is to simply remove safemalloc. Third-party tools
can provide the same functionality at a lower or comparable
performance cost.
The removal of safemalloc also allows a simplification of the
malloc wrappers, removing quite a bit of kludge: redefinition
of my_malloc, my_free and the removal of the unused second
argument of my_free. Since free() always check whether the
supplied pointer is null, redudant checks are also removed.
Also, this patch adds unit testing for my_malloc and moves
my_realloc implementation into the same file as the other
memory allocation primitives.
client/mysqldump.c:
Pass my_free directly as its signature is compatible with the
callback type -- which wasn't the case for free_table_ent.
2010-07-08 23:20:08 +02:00
|
|
|
my_free(to_free);
|
2007-02-26 17:44:55 +01:00
|
|
|
}
|
|
|
|
DBUG_ASSERT(tables_to_lock == NULL && tables_to_lock_count == 0);
|
Bug#13693012: SLAVE CRASHING ON INSERT STATEMENT WITH MERGE TABLE
PROBLEM: After WL 4144, when using MyISAM Merge tables, the routine
open_and_lock_tables will append to the list of tables to lock, the
base tables that make up the MERGE table. This has two side-effects in
replication:
1. On the master side, we log additional table maps for the base
tables, since they appear in the list of locked tables, even
though we don't really use them at the slave.
2. On the slave side, when opening a MERGE table while applying a
ROW event, additional tables are appended to the list of tables
to lock.
Side-effect #1 is not harmful. It's just that when using MyISAM Merge
tables a few table maps more may be logged.
Side-effect #2, is harmful, because the list rli->tables_to_lock is an
extended structure from TABLE_LIST in which the extra fields are
filled from the table maps that are processed. Since
open_and_lock_tables appends tables to the list after all table map
events have been processed we end up with entries without
replication/table map data on them. Thus when trying to access that
info for these extra tables, the server will crash.
SOLUTION: We fix side-effect #2 by making sure that we access the
replication part of the structure for those in the list that were
accounted for when processing the correspondent table map events. All
in all, we never go beyond rli->tables_to_lock_count.
We also deploy an assertion when clearing rli->tables_to_lock, making
sure that the base tables are not in the list anymore (were closed in
close_thread_tables).
2012-02-24 17:07:43 +01:00
|
|
|
DBUG_VOID_RETURN;
|
2007-02-26 17:44:55 +01:00
|
|
|
}
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
void Relay_log_info::slave_close_thread_tables(THD *thd)
|
|
|
|
{
|
Bug#13693012: SLAVE CRASHING ON INSERT STATEMENT WITH MERGE TABLE
PROBLEM: After WL 4144, when using MyISAM Merge tables, the routine
open_and_lock_tables will append to the list of tables to lock, the
base tables that make up the MERGE table. This has two side-effects in
replication:
1. On the master side, we log additional table maps for the base
tables, since they appear in the list of locked tables, even
though we don't really use them at the slave.
2. On the slave side, when opening a MERGE table while applying a
ROW event, additional tables are appended to the list of tables
to lock.
Side-effect #1 is not harmful. It's just that when using MyISAM Merge
tables a few table maps more may be logged.
Side-effect #2, is harmful, because the list rli->tables_to_lock is an
extended structure from TABLE_LIST in which the extra fields are
filled from the table maps that are processed. Since
open_and_lock_tables appends tables to the list after all table map
events have been processed we end up with entries without
replication/table map data on them. Thus when trying to access that
info for these extra tables, the server will crash.
SOLUTION: We fix side-effect #2 by making sure that we access the
replication part of the structure for those in the list that were
accounted for when processing the correspondent table map events. All
in all, we never go beyond rli->tables_to_lock_count.
We also deploy an assertion when clearing rli->tables_to_lock, making
sure that the base tables are not in the list anymore (were closed in
close_thread_tables).
2012-02-24 17:07:43 +01:00
|
|
|
DBUG_ENTER("Relay_log_info::slave_close_thread_tables(THD *thd)");
|
A pre-requisite patch for the fix for Bug#52044.
This patch also fixes Bug#55452 "SET PASSWORD is
replicated twice in RBR mode".
The goal of this patch is to remove the release of
metadata locks from close_thread_tables().
This is necessary to not mistakenly release
the locks in the course of a multi-step
operation that involves multiple close_thread_tables()
or close_tables_for_reopen().
On the same token, move statement commit outside
close_thread_tables().
Other cleanups:
Cleanup COM_FIELD_LIST.
Don't call close_thread_tables() in COM_SHUTDOWN -- there
are no open tables there that can be closed (we leave
the locked tables mode in THD destructor, and this
close_thread_tables() won't leave it anyway).
Make open_and_lock_tables() and open_and_lock_tables_derived()
call close_thread_tables() upon failure.
Remove the calls to close_thread_tables() that are now
unnecessary.
Simplify the back off condition in Open_table_context.
Streamline metadata lock handling in LOCK TABLES
implementation.
Add asserts to ensure correct life cycle of
statement transaction in a session.
Remove a piece of dead code that has also become redundant
after the fix for Bug 37521.
mysql-test/r/variables.result:
Update results: set @@autocommit and statement transaction/
prelocked mode.
mysql-test/r/view.result:
A harmless change in CHECK TABLE <view> status for a broken view.
If previously a failure to prelock all functions used in a view
would leave the connection in LTM_PRELOCKED mode, now we call
close_thread_tables() from open_and_lock_tables()
and leave prelocked mode, thus some check in mysql_admin_table() that
works only in prelocked/locked tables mode is no longer activated.
mysql-test/suite/rpl/r/rpl_row_implicit_commit_binlog.result:
Fixed Bug#55452 "SET PASSWORD is replicated twice in
RBR mode": extra binlog events are gone from the
binary log.
mysql-test/t/variables.test:
Add a test case: set autocommit and statement transaction/prelocked
mode.
sql/event_data_objects.cc:
Simplify code in Event_job_data::execute().
Move sp_head memory management to lex_end().
sql/event_db_repository.cc:
Move the release of metadata locks outside
close_thread_tables().
Make sure we call close_thread_tables() when
open_and_lock_tables() fails and remove extra
code from the events data dictionary.
Use close_mysql_tables(), a new internal
function to properly close mysql.* tables
in the data dictionary.
Contract Event_db_repository::drop_events_by_field,
drop_schema_events into one function.
When dropping all events in a schema,
make sure we don't mistakenly release all
locks acquired by DROP DATABASE. These
include locks on the database name
and the global intention exclusive
metadata lock.
sql/event_db_repository.h:
Function open_event_table() does not require an instance
of Event_db_repository.
sql/events.cc:
Use close_mysql_tables() instead of close_thread_tables()
to bootstrap events, since the latter no longer
releases metadata locks.
sql/ha_ndbcluster.cc:
- mysql_rm_table_part2 no longer releases
acquired metadata locks. Do it in the caller.
sql/ha_ndbcluster_binlog.cc:
Deploy the new protocol for closing thread
tables in run_query() and ndb_binlog_index
code.
sql/handler.cc:
Assert that we never call ha_commit_trans/
ha_rollback_trans in sub-statement, which
is now the case.
sql/handler.h:
Add an accessor to check whether THD_TRANS object
is empty (has no transaction started).
sql/log.cc:
Update a comment.
sql/log_event.cc:
Since now we commit/rollback statement transaction in
mysql_execute_command(), we need a mechanism to communicate
from Query_log_event::do_apply_event() to mysql_execute_command()
that the statement transaction should be rolled back, not committed.
Ideally it would be a virtual method of THD. I hesitate
to make THD a virtual base class in this already large patch.
Use a thd->variables.option_bits for now.
Remove a call to close_thread_tables() from the slave IO
thread. It doesn't open any tables, and the protocol
for closing thread tables is more complicated now.
Make sure we properly close thread tables, however,
in Load_data_log_event, which doesn't
follow the standard server execution procedure
with mysql_execute_command().
@todo: this piece should use Server_runnable
framework instead.
Remove an unnecessary call to mysql_unlock_tables().
sql/rpl_rli.cc:
Update Relay_log_info::slave_close_thread_tables()
to follow the new close protocol.
sql/set_var.cc:
Remove an unused header.
sql/slave.cc:
Remove an unnecessary call to
close_thread_tables().
sql/sp.cc:
Remove unnecessary calls to close_thread_tables()
from SP DDL implementation. The tables will
be closed by the caller, in mysql_execute_command().
When dropping all routines in a database, make sure
to not mistakenly drop all metadata locks acquired
so far, they include the scoped lock on the schema.
sql/sp_head.cc:
Correct the protocol that closes thread tables
in an SP instruction.
Clear lex->sphead before cleaning up lex
with lex_end to make sure that we don't
delete the sphead twice. It's considered
to be "cleaner" and more in line with
future changes than calling delete lex->sphead
in other places that cleanup the lex.
sql/sp_head.h:
When destroying m_lex_keeper of an instruction,
don't delete the sphead that all lex objects
share.
@todo: don't store a reference to routine's sp_head
instance in instruction's lex.
sql/sql_acl.cc:
Don't call close_thread_tables() where the caller will
do that for us.
Fix Bug#55452 "SET PASSWORD is replicated twice in RBR
mode" by disabling RBR replication in change_password()
function.
Use close_mysql_tables() in bootstrap and ACL reload
code to make sure we release all metadata locks.
sql/sql_base.cc:
This is the main part of the patch:
- remove manipulation with thd->transaction
and thd->mdl_context from close_thread_tables().
Now this function is only responsible for closing
tables, nothing else.
This is necessary to be able to easily use
close_thread_tables() in procedures, that
involve multiple open/close tables, which all
need to be protected continuously by metadata
locks.
Add asserts ensuring that TABLE object
is only used when is protected by a metadata lock.
Simplify the back off condition of Open_table_context,
we no longer need to look at the autocommit mode.
Make open_and_lock_tables() and open_normal_and_derived_tables()
close thread tables and release metadata locks acquired so-far
upon failure. This simplifies their usage.
Implement close_mysql_tables().
sql/sql_base.h:
Add declaration for close_mysql_tables().
sql/sql_class.cc:
Remove a piece of dead code that has also become redundant
after the fix for Bug 37521.
The code became dead when my_eof() was made a non-protocol method,
but a method that merely modifies the diagnostics area.
The code became redundant with the fix for Bug#37521, when
we started to cal close_thread_tables() before
Protocol::end_statement().
sql/sql_do.cc:
Do nothing in DO if inside a substatement
(the assert moved out of trans_rollback_stmt).
sql/sql_handler.cc:
Add comments.
sql/sql_insert.cc:
Remove dead code.
Release metadata locks explicitly at the
end of the delayed insert thread.
sql/sql_lex.cc:
Add destruction of lex->sphead to lex_end(),
lex "reset" method called at the end of each statement.
sql/sql_parse.cc:
Move close_thread_tables() and other related
cleanups to mysql_execute_command()
from dispatch_command(). This has become
possible after the fix for Bug#37521.
Mark federated SERVER statements as DDL.
Next step: make sure that we don't store
eof packet in the query cache, and move
the query cache code outside mysql_parse.
Brush up the code of COM_FIELD_LIST.
Remove unnecessary calls to close_thread_tables().
When killing a query, don't report "OK"
if it was a suicide.
sql/sql_parse.h:
Remove declaration of a function that is now static.
sql/sql_partition.cc:
Remove an unnecessary call to close_thread_tables().
sql/sql_plugin.cc:
open_and_lock_tables() will clean up
after itself after a failure.
Move close_thread_tables() above
end: label, and replace with close_mysql_tables(),
which will also release the metadata lock
on mysql.plugin.
sql/sql_prepare.cc:
Now that we no longer release locks in close_thread_tables()
statement prepare code has become more straightforward.
Remove the now redundant check for thd->killed() (used
only by the backup project) from Execute_server_runnable.
Reorder code to take into account that now mysql_execute_command()
performs lex->unit.cleanup() and close_thread_tables().
sql/sql_priv.h:
Add a new option to server options to interact
between the slave SQL thread and execution
framework (hack). @todo: use a virtual
method of class THD instead.
sql/sql_servers.cc:
Due to Bug 25705 replication of
DROP/CREATE/ALTER SERVER is broken.
Make sure at least we do not attempt to
replicate these statements using RBR,
as this violates the assert in close_mysql_tables().
sql/sql_table.cc:
Do not release metadata locks in mysql_rm_table_part2,
this is done by the caller.
Do not call close_thread_tables() in mysql_create_table(),
this is done by the caller.
Fix a bug in DROP TABLE under LOCK TABLES when,
upon error in wait_while_table_is_used() we would mistakenly
release the metadata lock on a non-dropped table.
Explicitly release metadata locks when doing an implicit
commit.
sql/sql_trigger.cc:
Now that we delete lex->sphead in lex_end(),
zero the trigger's sphead in lex after loading
the trigger, to avoid double deletion.
sql/sql_udf.cc:
Use close_mysql_tables() instead of close_thread_tables().
sql/sys_vars.cc:
Remove code added in scope of WL#4284 which would
break when we perform set @@session.autocommit along
with setting other variables and using tables or functions.
A test case added to variables.test.
sql/transaction.cc:
Add asserts.
sql/tztime.cc:
Use close_mysql_tables() rather than close_thread_tables().
2010-07-27 12:25:53 +02:00
|
|
|
thd->stmt_da->can_overwrite_status= TRUE;
|
|
|
|
thd->is_error() ? trans_rollback_stmt(thd) : trans_commit_stmt(thd);
|
|
|
|
thd->stmt_da->can_overwrite_status= FALSE;
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
close_thread_tables(thd);
|
A pre-requisite patch for the fix for Bug#52044.
This patch also fixes Bug#55452 "SET PASSWORD is
replicated twice in RBR mode".
The goal of this patch is to remove the release of
metadata locks from close_thread_tables().
This is necessary to not mistakenly release
the locks in the course of a multi-step
operation that involves multiple close_thread_tables()
or close_tables_for_reopen().
On the same token, move statement commit outside
close_thread_tables().
Other cleanups:
Cleanup COM_FIELD_LIST.
Don't call close_thread_tables() in COM_SHUTDOWN -- there
are no open tables there that can be closed (we leave
the locked tables mode in THD destructor, and this
close_thread_tables() won't leave it anyway).
Make open_and_lock_tables() and open_and_lock_tables_derived()
call close_thread_tables() upon failure.
Remove the calls to close_thread_tables() that are now
unnecessary.
Simplify the back off condition in Open_table_context.
Streamline metadata lock handling in LOCK TABLES
implementation.
Add asserts to ensure correct life cycle of
statement transaction in a session.
Remove a piece of dead code that has also become redundant
after the fix for Bug 37521.
mysql-test/r/variables.result:
Update results: set @@autocommit and statement transaction/
prelocked mode.
mysql-test/r/view.result:
A harmless change in CHECK TABLE <view> status for a broken view.
If previously a failure to prelock all functions used in a view
would leave the connection in LTM_PRELOCKED mode, now we call
close_thread_tables() from open_and_lock_tables()
and leave prelocked mode, thus some check in mysql_admin_table() that
works only in prelocked/locked tables mode is no longer activated.
mysql-test/suite/rpl/r/rpl_row_implicit_commit_binlog.result:
Fixed Bug#55452 "SET PASSWORD is replicated twice in
RBR mode": extra binlog events are gone from the
binary log.
mysql-test/t/variables.test:
Add a test case: set autocommit and statement transaction/prelocked
mode.
sql/event_data_objects.cc:
Simplify code in Event_job_data::execute().
Move sp_head memory management to lex_end().
sql/event_db_repository.cc:
Move the release of metadata locks outside
close_thread_tables().
Make sure we call close_thread_tables() when
open_and_lock_tables() fails and remove extra
code from the events data dictionary.
Use close_mysql_tables(), a new internal
function to properly close mysql.* tables
in the data dictionary.
Contract Event_db_repository::drop_events_by_field,
drop_schema_events into one function.
When dropping all events in a schema,
make sure we don't mistakenly release all
locks acquired by DROP DATABASE. These
include locks on the database name
and the global intention exclusive
metadata lock.
sql/event_db_repository.h:
Function open_event_table() does not require an instance
of Event_db_repository.
sql/events.cc:
Use close_mysql_tables() instead of close_thread_tables()
to bootstrap events, since the latter no longer
releases metadata locks.
sql/ha_ndbcluster.cc:
- mysql_rm_table_part2 no longer releases
acquired metadata locks. Do it in the caller.
sql/ha_ndbcluster_binlog.cc:
Deploy the new protocol for closing thread
tables in run_query() and ndb_binlog_index
code.
sql/handler.cc:
Assert that we never call ha_commit_trans/
ha_rollback_trans in sub-statement, which
is now the case.
sql/handler.h:
Add an accessor to check whether THD_TRANS object
is empty (has no transaction started).
sql/log.cc:
Update a comment.
sql/log_event.cc:
Since now we commit/rollback statement transaction in
mysql_execute_command(), we need a mechanism to communicate
from Query_log_event::do_apply_event() to mysql_execute_command()
that the statement transaction should be rolled back, not committed.
Ideally it would be a virtual method of THD. I hesitate
to make THD a virtual base class in this already large patch.
Use a thd->variables.option_bits for now.
Remove a call to close_thread_tables() from the slave IO
thread. It doesn't open any tables, and the protocol
for closing thread tables is more complicated now.
Make sure we properly close thread tables, however,
in Load_data_log_event, which doesn't
follow the standard server execution procedure
with mysql_execute_command().
@todo: this piece should use Server_runnable
framework instead.
Remove an unnecessary call to mysql_unlock_tables().
sql/rpl_rli.cc:
Update Relay_log_info::slave_close_thread_tables()
to follow the new close protocol.
sql/set_var.cc:
Remove an unused header.
sql/slave.cc:
Remove an unnecessary call to
close_thread_tables().
sql/sp.cc:
Remove unnecessary calls to close_thread_tables()
from SP DDL implementation. The tables will
be closed by the caller, in mysql_execute_command().
When dropping all routines in a database, make sure
to not mistakenly drop all metadata locks acquired
so far, they include the scoped lock on the schema.
sql/sp_head.cc:
Correct the protocol that closes thread tables
in an SP instruction.
Clear lex->sphead before cleaning up lex
with lex_end to make sure that we don't
delete the sphead twice. It's considered
to be "cleaner" and more in line with
future changes than calling delete lex->sphead
in other places that cleanup the lex.
sql/sp_head.h:
When destroying m_lex_keeper of an instruction,
don't delete the sphead that all lex objects
share.
@todo: don't store a reference to routine's sp_head
instance in instruction's lex.
sql/sql_acl.cc:
Don't call close_thread_tables() where the caller will
do that for us.
Fix Bug#55452 "SET PASSWORD is replicated twice in RBR
mode" by disabling RBR replication in change_password()
function.
Use close_mysql_tables() in bootstrap and ACL reload
code to make sure we release all metadata locks.
sql/sql_base.cc:
This is the main part of the patch:
- remove manipulation with thd->transaction
and thd->mdl_context from close_thread_tables().
Now this function is only responsible for closing
tables, nothing else.
This is necessary to be able to easily use
close_thread_tables() in procedures, that
involve multiple open/close tables, which all
need to be protected continuously by metadata
locks.
Add asserts ensuring that TABLE object
is only used when is protected by a metadata lock.
Simplify the back off condition of Open_table_context,
we no longer need to look at the autocommit mode.
Make open_and_lock_tables() and open_normal_and_derived_tables()
close thread tables and release metadata locks acquired so-far
upon failure. This simplifies their usage.
Implement close_mysql_tables().
sql/sql_base.h:
Add declaration for close_mysql_tables().
sql/sql_class.cc:
Remove a piece of dead code that has also become redundant
after the fix for Bug 37521.
The code became dead when my_eof() was made a non-protocol method,
but a method that merely modifies the diagnostics area.
The code became redundant with the fix for Bug#37521, when
we started to cal close_thread_tables() before
Protocol::end_statement().
sql/sql_do.cc:
Do nothing in DO if inside a substatement
(the assert moved out of trans_rollback_stmt).
sql/sql_handler.cc:
Add comments.
sql/sql_insert.cc:
Remove dead code.
Release metadata locks explicitly at the
end of the delayed insert thread.
sql/sql_lex.cc:
Add destruction of lex->sphead to lex_end(),
lex "reset" method called at the end of each statement.
sql/sql_parse.cc:
Move close_thread_tables() and other related
cleanups to mysql_execute_command()
from dispatch_command(). This has become
possible after the fix for Bug#37521.
Mark federated SERVER statements as DDL.
Next step: make sure that we don't store
eof packet in the query cache, and move
the query cache code outside mysql_parse.
Brush up the code of COM_FIELD_LIST.
Remove unnecessary calls to close_thread_tables().
When killing a query, don't report "OK"
if it was a suicide.
sql/sql_parse.h:
Remove declaration of a function that is now static.
sql/sql_partition.cc:
Remove an unnecessary call to close_thread_tables().
sql/sql_plugin.cc:
open_and_lock_tables() will clean up
after itself after a failure.
Move close_thread_tables() above
end: label, and replace with close_mysql_tables(),
which will also release the metadata lock
on mysql.plugin.
sql/sql_prepare.cc:
Now that we no longer release locks in close_thread_tables()
statement prepare code has become more straightforward.
Remove the now redundant check for thd->killed() (used
only by the backup project) from Execute_server_runnable.
Reorder code to take into account that now mysql_execute_command()
performs lex->unit.cleanup() and close_thread_tables().
sql/sql_priv.h:
Add a new option to server options to interact
between the slave SQL thread and execution
framework (hack). @todo: use a virtual
method of class THD instead.
sql/sql_servers.cc:
Due to Bug 25705 replication of
DROP/CREATE/ALTER SERVER is broken.
Make sure at least we do not attempt to
replicate these statements using RBR,
as this violates the assert in close_mysql_tables().
sql/sql_table.cc:
Do not release metadata locks in mysql_rm_table_part2,
this is done by the caller.
Do not call close_thread_tables() in mysql_create_table(),
this is done by the caller.
Fix a bug in DROP TABLE under LOCK TABLES when,
upon error in wait_while_table_is_used() we would mistakenly
release the metadata lock on a non-dropped table.
Explicitly release metadata locks when doing an implicit
commit.
sql/sql_trigger.cc:
Now that we delete lex->sphead in lex_end(),
zero the trigger's sphead in lex after loading
the trigger, to avoid double deletion.
sql/sql_udf.cc:
Use close_mysql_tables() instead of close_thread_tables().
sql/sys_vars.cc:
Remove code added in scope of WL#4284 which would
break when we perform set @@session.autocommit along
with setting other variables and using tables or functions.
A test case added to variables.test.
sql/transaction.cc:
Add asserts.
sql/tztime.cc:
Use close_mysql_tables() rather than close_thread_tables().
2010-07-27 12:25:53 +02:00
|
|
|
/*
|
|
|
|
- If inside a multi-statement transaction,
|
|
|
|
defer the release of metadata locks until the current
|
|
|
|
transaction is either committed or rolled back. This prevents
|
|
|
|
other statements from modifying the table for the entire
|
|
|
|
duration of this transaction. This provides commit ordering
|
|
|
|
and guarantees serializability across multiple transactions.
|
|
|
|
- If in autocommit mode, or outside a transactional context,
|
|
|
|
automatically release metadata locks of the current statement.
|
|
|
|
*/
|
|
|
|
if (! thd->in_multi_stmt_transaction_mode())
|
|
|
|
thd->mdl_context.release_transactional_locks();
|
2010-11-11 18:11:05 +01:00
|
|
|
else
|
|
|
|
thd->mdl_context.release_statement_locks();
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
clear_tables_to_lock();
|
Bug#13693012: SLAVE CRASHING ON INSERT STATEMENT WITH MERGE TABLE
PROBLEM: After WL 4144, when using MyISAM Merge tables, the routine
open_and_lock_tables will append to the list of tables to lock, the
base tables that make up the MERGE table. This has two side-effects in
replication:
1. On the master side, we log additional table maps for the base
tables, since they appear in the list of locked tables, even
though we don't really use them at the slave.
2. On the slave side, when opening a MERGE table while applying a
ROW event, additional tables are appended to the list of tables
to lock.
Side-effect #1 is not harmful. It's just that when using MyISAM Merge
tables a few table maps more may be logged.
Side-effect #2, is harmful, because the list rli->tables_to_lock is an
extended structure from TABLE_LIST in which the extra fields are
filled from the table maps that are processed. Since
open_and_lock_tables appends tables to the list after all table map
events have been processed we end up with entries without
replication/table map data on them. Thus when trying to access that
info for these extra tables, the server will crash.
SOLUTION: We fix side-effect #2 by making sure that we access the
replication part of the structure for those in the list that were
accounted for when processing the correspondent table map events. All
in all, we never go beyond rli->tables_to_lock_count.
We also deploy an assertion when clearing rli->tables_to_lock, making
sure that the base tables are not in the list anymore (were closed in
close_thread_tables).
2012-02-24 17:07:43 +01:00
|
|
|
DBUG_VOID_RETURN;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
2006-10-31 12:23:14 +01:00
|
|
|
#endif
|