Details
-
Bug
-
Status: Closed (View Workflow)
-
Critical
-
Resolution: Fixed
-
10.3(EOL), (15)
10.4(EOL), 10.5, 10.6, 10.7(EOL), 10.8(EOL), 10.9(EOL), 10.10(EOL), 10.11, 11.0(EOL), 11.1(EOL), 11.2(EOL), 11.3(EOL), 11.4, 11.6(EOL), 11.7(EOL)
Description
# mysqld options required for replay: --log-bin |
SET GLOBAL max_binlog_size=98304; # A valid and arbritary setting |
START TRANSACTION WITH CONSISTENT SNAPSHOT; |
Leads to:
|
10.11.0 6ebdd3013a18b01dbecec76b870810329eb76586 (Debug) |
safe_mutex: Found wrong usage of mutex 'LOCK_commit_ordered' and 'LOCK_global_system_variables'
|
And
# mysqld options required for replay: --log-bin |
START TRANSACTION WITH CONSISTENT SNAPSHOT; |
SET GLOBAL max_binlog_size=98304; # A valid and arbritary setting |
Leads to:
|
10.11.0 6ebdd3013a18b01dbecec76b870810329eb76586 (Debug) |
safe_mutex: Found wrong usage of mutex 'LOCK_global_system_variables' and 'LOCK_log'
|
Bug confirmed present in:
MariaDB: 10.3.37 (dbg), 10.4.27 (dbg), 10.5.18 (dbg), 10.6.10 (dbg), 10.7.6 (dbg), 10.8.5 (dbg), 10.9.3 (dbg), 10.10.2 (dbg), 10.11.0 (dbg)
Attachments
Issue Links
- is duplicated by
-
-
- Closed
-
- relates to
-
-
- Closed
-
Activity
Was able to catch a bit more info with this testcase
# mysqld options required for replay: --log-bin |
SET GLOBAL max_binlog_size=0; |
START TRANSACTION WITH CONSISTENT SNAPSHOT; |
|
10.11.0 6ebdd3013a18b01dbecec76b870810329eb76586 (Debug) |
safe_mutex: Found wrong usage of mutex 'LOCK_commit_ordered' and 'LOCK_global_system_variables'
|
Mutex currently locked (in reverse order):
|
LOCK_global_system_variables /test/10.11_dbg/sql/sql_plugin.cc line 3096
|
LOCK_commit_ordered /test/10.11_dbg/sql/handler.cc line 3007
|
Same source code line numbers for the testcase by elenst (with thanks!)
Also please try:
# mysqld options required for replay: --log-bin |
START TRANSACTION WITH CONSISTENT SNAPSHOT; |
FLUSH LOGS;
|
SET GLOBAL binlog_checksum=NONE; |
|
11.3.0 126157061b4376496c034a809ea4943e863d1465 (Debug) |
2023-12-07 15:08:21 0 [Note] /test/MD311023-mariadb-11.3.0-linux-x86_64-dbg/bin/mariadbd: ready for connections.
|
Version: '11.3.0-MariaDB-debug-log' socket: '/test/MD311023-mariadb-11.3.0-linux-x86_64-dbg/socket.sock' port: 10415 MariaDB Server
|
2023-12-07 15:08:21 6 [Note] Start binlog_dump to slave_server(2), pos(, 4), using_gtid(1), gtid('')
|
safe_mutex: Found wrong usage of mutex 'LOCK_global_system_variables' and 'LOCK_log'
|
Mutex currently locked (in reverse order):
|
LOCK_log /test/11.3_dbg/sql/log.cc line 12235
|
LOCK_global_system_variables /test/11.3_dbg/sql/sys_vars_shared.h line 56
|
# mysqld options required for replay: --log-bin |
SET GLOBAL binlog_checksum=DEFAULT; |
START TRANSACTION WITH CONSISTENT SNAPSHOT,READ WRITE; |
safe_mutex: Found wrong usage of mutex LOCK_commit_ordered and LOCK_global_system_variables
|
Elkin Hi! Would you have a moment to check how serious this issue is under the hood? Thank you
The same problem will probably also occur for max_binlog_total_size, which is available since 11.4.
Sure. It would be recognized as a known bug (this one) and ignored in tests.
knielsen Hi! Thank you for working on this. I see you put this in review but I cannot see an attached patch?
sergLet me know if you're going to review this week, otherwise I'll just take it back and push it with your suggestion to use mysql_mutex_record_order()
I was going to review it in about 2-3 weeks (note that the release is planned for the end of Jan)
Pushed to 10.6.
Requires another patch when merged to 11.4, see here: https://lists.mariadb.org/hyperkitty/list/developers@lists.mariadb.org/thread/LN35WETUEUHYUQNZXAKXLVU5EGSCQSVX/
Also with binlog_checksum, as in
--source include/have_innodb.inc--source include/have_log_bin.incThe problem is old though.