Re: [percona-group] innobackupex v2.03 fails with "log_copying_thread"

425 views
Skip to first unread message

Alexey Kopytov

unread,
Oct 5, 2012, 3:39:28 AM10/5/12
to percona-d...@googlegroups.com
Hi Dragan,

On 04.10.12 20:30, Dragan Prostran wrote:
> Issue: mysql backup job fails with innobackupex.
>
> Master info:
> OS: Ubuntu 10.04 LTS x86_64
> MySQL: ver 14.14 Distrib 5.1.63 for debian-linux-gnu (x86_64) using
> readline 6.1
> Innobackupex: percona-xtrabackup v2.0.3-470.lucid
>
> Backup command ran on master: sudo innobackupex --user=user
> --password=password --no-timestamp /backup/2012-10-03 2>2012-10-03.log
>
> Abridged contents of 2012-10-23.log (stderr of innobackupex during backup
> job):
>
[snip]
> xtrabackup: Error: log_copying_thread failed.
> 121003 19:33:12 innobackupex: All tables unlocked
> 121003 19:33:12 innobackupex: Connection to database server closed
>

Apparently something went wrong with the log copying thread, but it's
impossible to say what exactly, because the corresponding error message
is in the omitted part of the log.

Can you either post the full log, or at least just the lines produced by
"grep -i error"?

Best regards,
Alexey.

Alexey Kopytov

unread,
Oct 8, 2012, 10:24:37 AM10/8/12
to percona-d...@googlegroups.com
Hi Dragan,

On 06.10.12, 1:05, Dragan Prostran wrote:
> Hi Alexey,
>
> I believe this is the part you're interested in: the full log is 4000
> lines long, but the only instance of the string "failed" appears about
> 80 lines in. Here are thee first 85 lines of the log:
>
> <begin_partial_log_dump>
>
...
> xtrabackup: error: log block numbers mismatch:
> xtrabackup: error: expected log block no. 97070621, but got no. 97091093
> from the log file.
> xtrabackup: error: it looks like InnoDB log has wrapped around before
> xtrabackup could process all records due to either log copying being too
> slow, or log files being too small.
> xtrabackup: Error: xtrabackup_copy_logfile() failed.

Yes, that's the problem. Either log files are too small (it looks like
the default innodb_log_file_size is being used by the server) or writes
to the backup target device are too slow so the log copying activity in
XtraBackup is unable to catch up with the log writing activity in the
server. See discussion in
https://bugs.launchpad.net/percona-xtrabackup/+bug/1050355 for example.

Best regards,
Alexey.

Reply all
Reply to author
Forward
0 new messages