Mysql Keeps crashing

[timwiel]

Hi

The Zimbra MYSQL server keeps crashing on our ubuntu 8.04 LTS server.

The server has AMD Athlon(tm) XP 2000+ CPU, /opt/zimbra is on RAID1 (with two drives with 16GB free space), 1GB RAM.

This zimbra install has very little mail throughput into mailboxes - less than 100 messages per day would be my guess.

About 6 times within the last 2 months though the mysql server has had a serious crash requiring me to go through the mysql recovery from the wiki. It works everytime.

After the last time I updated from zcs-5.0.11_GA_2695.UBUNTU8.20081117052520 to 5.0.13_GA_2791.UBUNTU8.20090206174622 thinking it could be problems with a mysql bug or zimbra bug.

It has just crashed twice in the last 36 hours.

Various extractions from the logs give:

090428 14:14:30 [Note] /opt/zimbra/mysql/libexec/mysqld: ready for connections.
Version: '5.0.67-log'  socket: '/opt/zimbra/db/mysql.sock'  port: 7306  Source distribution
InnoDB: Database page corruption on disk or a failed
InnoDB: file read of page 52.
InnoDB: You may have to recover from a backup.
090428 15:41:15  InnoDB: Page dump in ascii and hex (16384 bytes):

and after the last crash recovery (with zimbra actually running) the mysql_error.log now reports after a zmcontrol stop && zmcontrol start:

090513 14:24:38 [Note] /opt/zimbra/mysql/libexec/mysqld: Normal shutdown

090513 14:24:38  InnoDB: Starting shutdown...
090513 14:24:41  InnoDB: Shutdown completed; log sequence number 0 6721454
090513 14:24:41 [Note] /opt/zimbra/mysql/libexec/mysqld: Shutdown complete

090513 14:24:41  mysqld ended

090513 14:26:34  mysqld started
090513 14:26:34 [Warning] option 'max_join_size': unsigned value 18446744073709551615 adjusted to 4294967295
090513 14:26:34 [Warning] option 'max_join_size': unsigned value 18446744073709551615 adjusted to 4294967295
090513 14:26:34  InnoDB: Error: page 4 log sequence number 0 32058696
InnoDB: is in the future! Current system log sequence number 0 6721454.
InnoDB: Your database may be corrupt or you may have copied the InnoDB
InnoDB: tablespace but not the InnoDB log files. See
InnoDB: http://dev.mysql.com/doc/refman/5.0/en/forcing-recovery.html
InnoDB: for more information.
090513 14:26:35  InnoDB: Started; log sequence number 0 6721454
090513 14:26:35 [Note] /opt/zimbra/mysql/libexec/mysqld: ready for connections.
Version: '5.0.67-log'  socket: '/opt/zimbra/db/mysql.sock'  port: 7306  Source distribution

Any help would be appreciated. Note we have three other clients running opensource zimbra editions without a hitch on Hardy and one client running a very active network edition. Never had these kinds of problems with theirs.

smartctl on the raid devices return no errors.

[phoenix]

Other than smartctl, have you run any other h/w diagnostics on this server? I'd suggest you run a memory check for at least 24 hrs (preferably 48hrs) on this machine and some HD tests, inexplicable behaviour like this would lead me to look at RAM problems.

Just for the record, what's the specification of this server?