[debian-mysql] MySQL server random crashes (Lenny)

[azurIt]

Well, I must say, that we have few signes which points to possible HW problems with hard drives or RAID controller.

We are using AMCC 9690SA RAID controller with four 500 GB drives (RAID 5). The controller is detecting random drive timeouts on all 4 drives quite often:
0009 Drive timeout detected
A drive has failed to respond to a command from a 3ware RAID controller within the allowed time limit (20 secs.). After sending this error message, the controller will attempt to recover the drive by sending a reset to that drive and retrying the failed command.
Possible causes of drive time-outs (also known as ATA-Port time-outs) include a bad or intermittent disk drive, power cable or interface cable.

But all drives looks ok, no reallocated sectors, SMART tests passes etc.

I'm also sometimes seeing messages like this in dmesg:
[1948128.486702] grsec: From x.x.x.x: Invalid alignment/Bus error occurred at 000004df000008a7 in /usr/sbin/apache2[apache2:2215] uid/euid:1247/1247 gid/egid:100/100, parent /usr/sbin/apache2[apache2:31879] uid/euid:0/0 gid/egid:0/0 

I was just hoping it's not failing RAID controller.. Colud it be related to MySQL crashes ?


>-----Pôvodná správa-----
>Od: Norbert Tretkowski [mailto:norbert at tretkowski.de]
>Komu: azurIt <azurit at pobox.sk>
>Predmet: Re: [debian-mysql] MySQL server random crashes (Lenny)
>
>
>Am Donnerstag, den 16.09.2010, 12:48 +0200 schrieb azurIt:
>> we are experiencing random crashes of MySQL server (package
>> mysql-server-5.0 version 5.0.51a-24+lenny4). I'm sending part of
>> syslog log + my.cnf as an attachment. Server is crashing about once
>> per day.
>
>Are you sure that the hardware is ok?
>
>Norbert