Bug #61579 | replication failure while innodb_autoinc_lock_mode=1 and using trigger | ||
---|---|---|---|
Submitted: | 21 Jun 2011 11:27 | Modified: | 17 May 2012 18:13 |
Reporter: | Hao Wu | Email Updates: | |
Status: | Closed | Impact on me: | |
Category: | MySQL Server: Replication | Severity: | S2 (Serious) |
Version: | 5.1.57, 5.5.14 | OS: | Linux (CentOS 5.5) |
Assigned to: | CPU Architecture: | Any |
[21 Jun 2011 11:27]
Hao Wu
[21 Jun 2011 11:29]
Hao Wu
script to reproduce the bug
Attachment: bug_61579.pl (application/octet-stream, text), 434 bytes.
[21 Jun 2011 11:29]
Hao Wu
script to reproduce the bug
Attachment: bug_61579.pl (application/octet-stream, text), 434 bytes.
[21 Jun 2011 11:32]
Hao Wu
Sorry for duplicate upload. It can be reproduced on 5.6.2-m5 too.
[22 Jun 2011 12:56]
MySQL Verification Team
Could you please provide the my.cnf files for master and slave. Thanks.
[23 Jun 2011 1:43]
Hao Wu
cnf of master
Attachment: my.master.cnf (application/octet-stream, text), 1.72 KiB.
[23 Jun 2011 1:44]
Hao Wu
cnf of slave
Attachment: my.slave.cnf (application/octet-stream, text), 1.76 KiB.
[13 Jul 2011 19:56]
Sveta Smirnova
Thank you for the feedback. Verified as described.
[17 May 2012 18:13]
John Russell
Added to changelog for 5.1.64, 5.5.26, 5.6.6: For an InnoDB table with a trigger, under the setting innodb_autoinc_lock_mode=1, sometimes auto-increment values could be interleaved when inserting into the table from two sessions concurrently. The sequence of auto-increment values could vary depending on timing, leading to data inconsistency in systems using replication.