Bug #79574 InnoDB: unable to purge a record

• Submitted: 9 Dec 2015 12:15
• Modified: 6 Jul 2018 14:57
• Reporter: Michael Brunnbauer
• Status: Can't repeat
• Category: MySQL Server: InnoDB storage engine
• Severity: S2 (Serious)
• Version: 5.6.27
• OS: Linux (Kernel 3.4.109, x86_64, glibc-2.19)
• CPU Architecture: Any

[9 Dec 2015 12:15] Michael Brunnbauer

Description:
2015-11-16 11:51:25 1844 [Note] /usr/local/mysql/bin/mysqld: ready for connections.
Version: '5.6.27'  socket: '/tmp/mysql.sock'  port: 3306  MySQL Community Server (GPL)
2015-12-08 18:50:34 7f701932a700  InnoDB: unable to purge a record
InnoDB: tuple DATA TUPLE: 2 fields;
 0: SQL NULL;
 1: len 40; hex 7777772e6266792e7777772e6168742e7777772e38326b2e7777772e3662652e393434386c2e7077; asc www.bfy.www.aht.www.82k.www.6be.9448l.pw;;

InnoDB: record PHYSICAL RECORD: n_fields 2; compact format; info bits 32
 0: SQL NULL;
 1: len 30; hex 7777772e6266792e7777772e6168742e7777772e38326b2e7777772e3662; asc www.bfy.www.aht.www.82k.www.6b; (total 40 bytes);

space 1016 offset 7829092 (1 records, index id 240)
InnoDB: Submit a detailed bug report to http://bugs.mysql.com

How to repeat:
No idea how to reproduce. The error message said to submit a bug report here and this is what I did. Would be grateful for advice on how to proceed (check table?).

[6 Jul 2018 14:57] MySQL Verification Team

Hi,

Thank you for submitting this report.

However, we do need a test case, so that we could repeat the problem and then fix it. If there is no test case, what might help would be a full stack trace with values of all local variables in all of the functions within the stack calls.

Regarding what might be a cause of this, look at the health of your I/O system, with stress test on all of the partitions. Also, this can happen if you have got some filesystem that has little of free space left. Also, a change in ownerships and umask could cause this problem.