Commit graph

1424 commits

Author SHA1 Message Date
Marko Mäkelä
9a46d97149 MDEV-15333 MariaDB (still) slow start
This performance regression was introduced in the MariaDB 10.1
file format incompatibility bug fix MDEV-11623 (MariaDB 10.1.21
and MariaDB 10.2.4) and partially fixed in MariaDB 10.1.25 in
MDEV-12610 without adding a regression test case.

On a normal startup (without crash recovery), InnoDB should not read
every .ibd data file, because this is slow. Like in MySQL, for now,
InnoDB will still open every data file (without reading), and it
will read every .ibd file for which an .isl file exists, or the
DATA DIRECTORY attribute has been specified for the table.

The test case shuts down InnoDB, moves data files, replaces them
with garbage, and then restarts InnoDB, expecting no messages to
be issued for the garbage files. (Some messages will for now be
issued for the table that uses the DATA DIRECTORY attribute.)
Finally, the test shuts down the server, restores the old data files,
and restarts again to drop the tables.

fil_open_single_table_tablespace(): Remove the condition on flags,
and only call fsp_flags_try_adjust() if validate==true
(reading the first page has been requested). The only caller with
validate==false is at server startup when we are processing all
records from SYS_TABLES. The flags passed to this function are
actually derived from SYS_TABLES.TYPE and SYS_TABLES.N_COLS,
and there never was any problem with SYS_TABLES in MariaDB 10.1.
The problem that MDEV-11623 was that incorrect tablespace flags
were computed and written to FSP_SPACE_FLAGS.
2018-02-17 14:20:33 +02:00
Marko Mäkelä
2202afd541 Merge 10.0 into 10.1 2018-02-13 14:32:17 +02:00
Marko Mäkelä
c051eaba46 MDEV-14988 innodb_read_only tries to modify files if transactions were recovered in COMMITTED state
lock_trx_release_locks(): Relax a debug assertion to allow
recovered TRX_STATE_COMMITTED_IN_MEMORY transactions.

trx_commit_in_memory(): Add DEBUG_SYNC instrumentation.

trx_undo_insert_cleanup(): Skip persistent changes if innodb_read_only
is set. This should only happen when a recovered committed transaction
would be cleaned up at shutdown.
2018-02-13 14:29:32 +02:00
Marko Mäkelä
871f2a6ee2 Merge 10.0 into 10.1 2018-02-08 13:29:08 +02:00
Marko Mäkelä
9216a4f69f Make the test innodb.recovery_shutdown more robust
Before killing the server, we have to FLUSH TABLES in order to
prevent the corruption of any MyISAM system tables.
2018-02-08 13:26:44 +02:00
Marko Mäkelä
5421e3aee7 MDEV-15249 Crash in MVCC read after IMPORT TABLESPACE
PageConverter::adjust_cluster_record(): Instead of writing
the invalid value DB_ROLL_PTR=0, write a value that indicates
a fresh insert, that is, prevents the DB_ROLL_PTR from being
dereferenced in any circumstances.

It can be argued that IMPORT TABLESPACE should actually
update the dict_index_t::trx_id to prevent older transactions
from accessing the table, similar to what I did on table
rebuild in MySQL 5.6.6 in
03f81a55f2
2018-02-08 12:59:32 +02:00
Marko Mäkelä
3f42529a6f Merge 10.0 into 10.1 2018-02-05 09:25:33 +02:00
Marko Mäkelä
cb5374801e MDEV-15202 innodb.log_file_size failed in buildbot
Suppress some messages that are emitted rarely
(when the FIL_PAGE_FILE_FLUSH_LSN of the first page of ibdata1
does not match the latest redo log checkpoint).
2018-02-05 09:23:36 +02:00
Sergei Golubchik
d4df7bc9b1 Merge branch 'github/10.0' into 10.1 2018-02-02 10:09:44 +01:00
Marko Mäkelä
b76881a23c Do not SET DEBUG_DBUG=-d,... in tests
To disable debug instrumentation, save and restore the original value
of the variable DEBUG_DBUG. Assigning -d,... will enable the output of
a lot of unrelated DBUG messages to the server error log.
2018-01-29 16:39:54 +02:00
Marko Mäkelä
a5fcced7d1 Merge 5.5 into 10.0 2018-01-29 16:32:59 +02:00
Marko Mäkelä
547ec8ce27 Do not SET DEBUG_DBUG=-d,... in tests
To disable debug instrumentation, save and restore the original value
of the variable DEBUG_DBUG. Assigning -d,... will enable the output of
a lot of unrelated DBUG messages to the server error log.
2018-01-29 16:25:26 +02:00
Vicențiu Ciorbaru
d69d488b8c Remove innodb.test "keep away" comment
This was a leftover post 5.5->10.0 merge. It should've been deleted
there.
2018-01-24 17:55:26 +02:00
Vladislav Vaintroub
524221e7a3 try to fix [Warning] InnoDB: innodb_open_files=20 is exceeded (20) files stay open)
on Appveyor
2018-01-24 11:23:44 +00:00
Vicențiu Ciorbaru
d833bb65d5 Merge remote-tracking branch '5.5' into 10.0 2018-01-24 12:29:31 +02:00
Marko Mäkelä
906ce0962d MDEV-7049 MySQL#74585 - InnoDB: Failing assertion: *mbmaxlen < 5 in file ha_innodb.cc line 1904
InnoDB limited the maximum number of bytes per character to 4.
But, the filename character set that was introduced in MySQL 5.1
uses up to 5 bytes per character.

To allow InnoDB tables to be created with wider characters, let
us split the mbminmaxlen fields into mbminlen, mbmaxlen, and increase
the limit to 7 bytes per character. This will increase the payload size
of dtype_t and dict_col_t by one bit. The storage size will be unchanged
(54 bits and 77 bits will use the same number of bytes as the
previous sizes 53 and 76 bits).
2018-01-22 11:18:10 +02:00
Marko Mäkelä
4794e5b091 Fix a test that always failed on --embedded 2018-01-15 16:19:46 +02:00
Marko Mäkelä
c15b3d2d41 Merge 10.0 into 10.1 2018-01-11 10:44:05 +02:00
Marko Mäkelä
4c1479545d Merge 5.5 into 10.0 2018-01-11 10:16:52 +02:00
Marko Mäkelä
a9c55c0059 MDEV-13814 Extra logging when innodb_log_archive=ON
Backport the fix from 10.0.33 to 5.5, in case someone compiles XtraDB
with -DUNIV_LOG_ARCHIVE
2018-01-10 10:21:52 +02:00
Marko Mäkelä
a408e881cf MDEV-14174 crash on start with innodb-track-changed-pages
The XtraDB option innodb_track_changed_pages causes
the function log_group_read_log_seg() to be invoked
even when recv_sys==NULL, leading to the SIGSEGV.

This regression was caused by
MDEV-11027 InnoDB log recovery is too noisy
2018-01-10 09:27:01 +02:00
Jan Lindström
07aa985979 MDEV-14776: InnoDB Monitor output generated by specific error is flooding error logs
innodb/buf_LRU_get_free_block
	Add debug instrumentation to produce error message about
	no free pages. Print error message only once and do not
	enable innodb monitor.

xtradb/buf_LRU_get_free_block
	Add debug instrumentation to produce error message about
	no free pages. Print error message only once and do not
	enable innodb monitor. Remove code that does not seem to
	be used.

innodb-lru-force-no-free-page.test
	New test case to force produce desired error message.
2018-01-09 12:48:31 +02:00
Marko Mäkelä
29b6e809a9 Merge 10.0 into 10.1 2018-01-08 14:51:20 +02:00
Marko Mäkelä
c903ba2f1e MDEV-13205 InnoDB: Failing assertion: !dict_index_is_online_ddl(index) upon ALTER TABLE
dict_foreign_find_index(): Ignore incompletely created indexes.
After a failed ADD UNIQUE INDEX, an incompletely created index
could be left behind until the next ALTER TABLE statement.
2018-01-08 14:26:55 +02:00
Vicențiu Ciorbaru
d1c2cd30b7 Merge remote-tracking branch '10.0' into 10.1 2017-12-27 17:50:39 +02:00
Sergey Vojtovich
4b8cd4536a MDEV-13626 Merge InnoDB test cases from MySQL 5.7
Coverage for temporary tables modifications in read-only transactions.
Introduced in 5.7 by 325cdf426
2017-12-22 14:03:25 +04:00
Marko Mäkelä
9d7c0882fa Merge 10.0 into 10.1 2017-12-21 17:25:51 +02:00
Marko Mäkelä
0202e47274 MDEV-12827 Assertion failure when reporting duplicate key error in online table rebuild
row_log_table_apply_insert_low(), row_log_table_apply_update():
When reporting the error_key_num, only count the clustered index
if it corresponds to a key in the SQL layer.

The assertion failure was probably introduced by the (incomplete)
MySQL 5.6.28 bug fix
Bug  THE BOGUS DUPLICATE KEY ERROR IN ONLINE DDL
WITH INCORRECT KEY NAME
which we are improving.

Side note: the fix was incorrectly merged to MySQL 5.7.10;
incorrect key names will continue to be reported in MySQL 5.7.
2017-12-21 17:19:13 +02:00
Vicențiu Ciorbaru
e3d89652e5 Merge branch '10.0' into 10.1 2017-12-20 13:30:05 +02:00
Alexander Barkov
cb121a047b An after-fix for MDEV-14008 Assertion failing: `!is_set() || (m_status == DA_OK_BULK && is_bulk_op())
Fixing an additional failure discovered after a merge to 10.2
2017-12-20 13:49:27 +04:00
Marko Mäkelä
09c5bbf471 Merge 10.0 into 10.1 2017-12-18 20:05:50 +02:00
Alexander Barkov
c1e5fef05d MDEV-14008 Assertion failing: `!is_set() || (m_status == DA_OK_BULK && is_bulk_op()) 2017-12-18 11:25:38 +04:00
Marko Mäkelä
46305b006b Merge 10.0 into 10.1 2017-12-13 19:12:16 +02:00
Marko Mäkelä
b46fa627ca MDEV-12352 InnoDB shutdown should not be blocked by a large transaction rollback
row_undo_step(), trx_rollback_active(): Abort the rollback of a
recovered ordinary transaction if fast shutdown has been initiated.

trx_rollback_resurrected(): Convert an aborted-rollback transaction
into a fake XA PREPARE transaction, so that fast shutdown can proceed.
2017-12-13 18:02:09 +02:00
Marko Mäkelä
6559ba71a5 MDEV-13797 InnoDB may hang if shutdown is initiated soon after startup while rolling back recovered incomplete transactions
trx_rollback_resurrected(): If shutdown was initiated, fake all
remaining active transactions to XA PREPARE state, so that shutdown
can proceed. Also, make the parameter "all" an output that will be
assigned to FALSE in this case.

trx_rollback_or_clean_recovered(): Remove the shutdown check
(it was moved to trx_rollback_resurrected()).

trx_undo_free_prepared(): Relax assertions.
2017-12-13 16:30:38 +02:00
Marko Mäkelä
b93a87f186 Try to prevent sporadic test failures
On Windows, sometimes more files stay open:

[Warning] InnoDB: innodb_open_files=13 is exceeded (15 files stay open)
2017-12-13 12:52:28 +02:00
Jan Lindström
d7349e204b MDEV-9663: InnoDB assertion failure: *cursor->index->name == TEMP_INDEX_PREFIX
MariaDB adjustments to test case innodb-replace-debug and
add missing instrumentation to row0ins.cc.

MariaDB 10.1 does not seem to be affected.
2017-11-16 14:05:49 +02:00
Jan Lindström
eeec64d75e MDEV-9663: InnoDB assertion failure: *cursor->index->name == TEMP_INDEX_PREFIX
Imported missing test case from MySQL 5.7 for

commit 25781c154396dbbc21023786aa3be070057d6999
Author: Annamalai Gurusami <annamalai.gurusami@oracle.com>
Date:   Mon Feb 24 14:00:03 2014 +0530

    Bug  ASSERTION: *CURSOR->INDEX->NAME == TEMP_INDEX_PREFIX
2017-11-16 13:58:09 +02:00
Jan Lindström
93326ef051 MDEV-9663: InnoDB assertion failure: *cursor->index->name == TEMP_INDEX_PREFIX
MariaDB adjustments to test case innodb-replace-debug.

MariaDB 10.0 does not seem to be affected.
2017-11-16 13:21:07 +02:00
Jan Lindström
923ea5dbf6 MDEV-9663: InnoDB assertion failure: *cursor->index->name == TEMP_INDEX_PREFIX
Imported missing test case from MySQL 5.7 for

commit 25781c154396dbbc21023786aa3be070057d6999
Author: Annamalai Gurusami <annamalai.gurusami@oracle.com>
Date:   Mon Feb 24 14:00:03 2014 +0530

    Bug  ASSERTION: *CURSOR->INDEX->NAME == TEMP_INDEX_PREFIX
2017-11-16 13:18:22 +02:00
Jan Lindström
c44ece7342 MDEV-9663: InnoDB assertion failure: *cursor->index->name == TEMP_INDEX_PREFIX
MariaDB adjustments to test case innodb-replace-debug.
2017-11-16 12:56:54 +02:00
Jan Lindström
f7b110bdc1 MDEV-9663: InnoDB assertion failure: *cursor->index->name == TEMP_INDEX_PREFIX
Imported missing test case from MySQL 5.7 for

commit 25781c154396dbbc21023786aa3be070057d6999
Author: Annamalai Gurusami <annamalai.gurusami@oracle.com>
Date:   Mon Feb 24 14:00:03 2014 +0530

    Bug  ASSERTION: *CURSOR->INDEX->NAME == TEMP_INDEX_PREFIX

MariaDB 5.5 does not seem to be affected.
2017-11-16 12:39:41 +02:00
Marko Mäkelä
5691109689 Merge 10.0 into 10.1 2017-11-06 18:10:23 +02:00
Marko Mäkelä
6a524fcfdd MDEV-14140 IMPORT TABLESPACE must not go beyond FSP_FREE_LIMIT
ibuf_check_bitmap_on_import(): Only access the pages that
are below FSP_FREE_LIMIT. It is possible that especially with
ROW_FORMAT=COMPRESSED, the FSP_SIZE will be much bigger than
the FSP_FREE_LIMIT, and the bitmap pages (page_size*N, 1+page_size*N)
are filled with zero bytes.

buf_page_is_corrupted(), buf_page_io_complete(): Make the
fault injection compatible with MariaDB 10.2.

Backport the IMPORT tests from 10.2.
2017-11-06 14:55:34 +02:00
Marko Mäkelä
88edb1b3ed MDEV-14219 Allow online table rebuild when encryption or compression parameters change
When MariaDB 10.1.0 introduced table options for encryption and
compression, it unnecessarily changed
ha_innobase::check_if_supported_inplace_alter() so that ALGORITHM=COPY
is forced when these parameters differ.

A better solution is to move the check to innobase_need_rebuild().
In that way, the ALGORITHM=INPLACE interface (yes, the syntax is
very misleading) can be used for rebuilding the table much more
efficiently, with merge sort, with no undo logging, and allowing
concurrent DML operations.
2017-10-31 09:10:25 +02:00
Marko Mäkelä
db203d7471 Merge 10.0 into 10.1 2017-10-24 20:36:03 +03:00
Marko Mäkelä
44ed243522 Merge 5.5 into 10.0 2017-10-24 19:25:19 +03:00
Marko Mäkelä
439a7c994a MDEV-14051 'Undo log record is too big.' error occurring in very narrow range of string lengths
InnoDB was writing unnecessary information to the
update undo log records. Most notably, if an indexed column is updated,
the old value of the column would be logged twice: first as part of
the update vector, and then another time because it is an indexed column.

Because the InnoDB undo log record must fit in a single page,
this would cause unnecessary failure of certain updates.

Even after this fix, InnoDB still seems to be unnecessarily logging
indexed column values for non-updated columns. It seems that non-updated
secondary index columns only need to be logged when a PRIMARY KEY
column is updated. To reduce risk, we are not fixing this remaining flaw
in GA versions.

trx_undo_page_report_modify(): Log updated indexed columns only once.
2017-10-24 17:00:04 +03:00
Sergei Golubchik
9d2e2d7533 Merge branch '10.0' into 10.1 2017-10-22 13:03:41 +02:00
Marko Mäkelä
dfd010ef90 MDEV-14086 Setting innodb_buffer_pool_load_now or innodb_buffer_load_abort will crash if innodb_read_only
When MySQL 5.6.10 introduced innodb_read_only mode, it skipped the
creation of the InnoDB buffer pool dump/restore subsystem in that mode.
Attempts to set the variable innodb_buf_pool_dump_now would have
no effect in innodb_read_only mode, but the corresponding condition
was forgotten in from the other two update functions.

MySQL 5.7.20 would fix the innodb_buffer_pool_load_now,
but not innodb_buffer_pool_load_abort. Let us fix both in MariaDB.
2017-10-18 06:35:11 +03:00