Bug #78195 Bug #77684 leads to practically unrecoverable slave

• Submitted: 25 Aug 2015 0:39
• Modified: 25 Aug 2015 8:42
• Reporter: Sveta Smirnova (OCA)
• Status: Verified
• Category: MySQL Server: Replication
• Severity: S1 (Critical)
• Version: 5.6.26, 5.7.8
• OS: Any
• CPU Architecture: Any

[25 Aug 2015 0:39] Sveta Smirnova

Description:
Bug #77684 not only corrupts replication, but makes it practically unrecoverable.

How to repeat:
Follow "How to repeat" instruction from Bug #77684, then try to skip failing transaction as described at http://dev.mysql.com/doc/refman/5.6/en/replication-gtids-failover.html#replication-gtids-f...

[25 Aug 2015 0:39] Sveta Smirnova

test case for MTR

Attachment: rpl_issue56487.test (application/octet-stream, text), 962 bytes.

[25 Aug 2015 0:39] Sveta Smirnova

option file for master

Attachment: rpl_issue56487-master.opt (application/octet-stream, text), 62 bytes.

[25 Aug 2015 0:40] Sveta Smirnova

option file for slave

Attachment: rpl_issue56487-slave.opt (application/octet-stream, text), 124 bytes.

[25 Aug 2015 0:40] Sveta Smirnova

Workaround: manually edit relay-log.info and change both (!) master and relay log position, then restart slave's mysqld

[25 Aug 2015 8:42] MySQL Verification Team

Hello Sveta,

Thank you for the report and test case.
Verified as described on 5.6.26 and 5.7.8.

Thanks,
Umesh

[25 Aug 2015 8:43] MySQL Verification Team

test results

Attachment: 78195.results (application/octet-stream, text), 16.87 KiB.