Got an error from an unknown thread

[Greg Hazel]

Happened after an unclean mysqld restart, when it decided to check a table.

120716 18:57:08 [ERROR] Got an error from unknown thread, /home/jenkins/workspace/percona-server-5.5-rpms/label_exp/centos6-64/target/BUILD/Percona-Server-5.5.24-rel26.0/Percona-Server-5.5.24-rel26.0/storage/myisam/ha_myisam.cc:914

[Raghavendra D Prabhu (raghavendra-prabhu)]

The error is from mi_mark_crashed(file) in ha_myisam::check, when it calls _mi_report_crashed, and the thread which has table open is unknown. But to know more, you will need to provide more context to this error message from error log.

[Greg Hazel (ghazelgmail.com)]

There isn't really more context, except that the server was starting and checking a MyISAM table.

[Alexey Kopytov (akopytov)]

"Unknown" in this context means that there was no user connection associated with the thread where MyISAM table corruption was detected. Most likely a system table corruption was detected on a server startup. I'm not sure what the exact problem is. Setting to Invalid, feel free to reopen if I'm missing something.

[Andrejs (dubis-lv)]

Got the same error for another Percona version

130723 22:04:31 [ERROR] /usr/local/percona/bin/mysqld: Table './mm/DispatcherLog5' is marked as crashed and should be repaired
130723 22:04:31 [Warning] Checking table: './mm/DispatcherLog5'
130724 0:29:28 [ERROR] Got an error from unknown thread, /home/nacho/src/release-5.5.30-30.2/Percona-Server-5.5.30-rel30.2/storage/myisam/ha_myisam.cc:915

After mysql server crash and restart I have one client only (mysqldump).

Mysqldump was called with following parameters:

mysqldump --defaults-file=/etc/mysql-dbp/my.cnf --result-file=/bkp/dbp-mysql/schema//mm.sql --force --no-data --no-create-info=0 mm

[Andrejs (dubis-lv)]

I was tried to repair the table by myisamchk and got segmentation fault

[root@bkp-cc mm]# time /usr/local/percona/bin/myisamchk -r -f -v DispatcherLog5.MYI
- recovering (with sort) MyISAM-table 'DispatcherLog5.MYI'
Data records: 98618930
- Fixing index 1
  - Searching for keys, allocating buffer for 107113 keys
Segmentation fault

real 12m30.631s
user 2m29.222s
sys 1m9.357s

[root@bkp-cc mm]# ls -l DispatcherLog5.*
-rw-rw---- 1 mysql mysql 8916 Jun 26 2007 DispatcherLog5.frm
-rw-rw---- 1 mysql mysql 20750372995 Jul 22 10:14 DispatcherLog5.MYD
-rw-rw---- 1 mysql mysql 25949122560 Jul 24 11:12 DispatcherLog5.MYI
-rw-r----- 1 root root 10680532992 Jul 24 11:24 DispatcherLog5.TMD

[Valerii Kravchuk (valerii-kravchuk)]

If you are still affected by this bug, please, share your my.cnf file content. I wonder can it be just some kind of "out of memory" condition as table seems big enough.

[Valerii Kravchuk (valerii-kravchuk)]

Disk full is other possible reason, entire error log (or few more rows after the messages shared here at least) may help to find out more. See http://bugs.mysql.com/bug.php?id=69926

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PS-2776