After a long search we found out that the firmware & BIOS of raid
adaptor was to blame !
It was the ServeRaid-4L scsi controller of a IBM Netfinity. This card is
a new product. After the upgrade of BIOS and Firmware with version 4.50
(was 4.40) the server seems stable and some other strange behaviours I
did not investigate so far, dissapeared. Realy if you have a
serveRaid-4L-controler with firmware version 4.40, you should upgrade
immediately. I contacted IBM 2 weeks ago and they said the knew nothing
about problems with the new serveRaid adapter. This has cost a lot of time !
How did I dare to blame mysql and linux!
No doubt anymore, Mysql and linux are great !
Herald
Herald van der Breggen wrote:
> Hello Sinsa,
>
> I just changed the kernel to a non-smp-kernel, to prevent the problem
> you described. I checked: the kernel sees only 1 processor.
>
> Within an hour, mysql did it again ! The server load suddenly goes
> from 0.3 to over 30. The swap growed from about 6 Mb to over 110 Mb. I
> was just in time to stop mysqld.
>
> So, there is another thing than the smp bug as you described !
>
> I'll try mysql 3.22 and I will keep you informed.
>
> Regards,
> Herald
>
> Sinisa Milivojevic wrote:
>
>> Herald van der Breggen writes:
>> > Hello Noor,
>> > > I am struggling with a similar problem on a linux based system.
>> It's a > Netfinity 4500R, dual pentium, RAID-5 with a rather new
>> ServeRaid-4L > scsi controler.
>> > > On certain moments, mostly in the peak hours, mysql suddenly
>> behaves > weird and the load increases very very quick. The swap
>> used is also > increasing very quick (was that also the case on your
>> system ?).
>> > > We use mysql in combination with Apache::Session which uses
>> GET_LOCK. We > saw that no lock could be obtained, which could be
>> the cause for an > increase og threads being used.
>> > > At the moment the wevserver and mysql server are restarted
>> automically > when the number of threads is higher than 60 (by a
>> cron job), which is a > clumbsy solution of course....
>> > > I have seen this problem with 3.23.29-beta, 3.23.30-gamma and
>> 3.23.31.
>> > I consider downgrading to 3.22.x :-/
>> > > > Herald
>> > >
>>
>> Hi!
>>
>> The above is an already reported problem that occurs only on SMP Linux
>> on the high load.
>>
>> A solution is to try kernel 2.4, glibc 2.2 (when stable) and to build
>> properly MySQL on that setup.
>>
>>
>> Regards,
>>
>> Sinisa
>>
>> ____ __ _____ _____ ___ == MySQL AB
>> /*/\*\/\*\ /*/ \*\ /*/ \*\ |*| Sinisa Milivojevic
>> /*/ /*/ /*/ \*\_ |*| |*||*| mailto:[EMAIL PROTECTED]
>> /*/ /*/ /*/\*\/*/ \*\|*| |*||*| Larnaka, Cyprus
>> /*/ /*/ /*/\*\_/*/ \*\_/*/ |*|____
>> ^^^^^^^^^^^^/*/^^^^^^^^^^^\*\^^^^^^^^^^^
>> /*/ \*\ Developers Team
>>
>> ---------------------------------------------------------------------
>> Before posting, please check:
>> http://www.mysql.com/manual.php (the manual)
>> http://lists.mysql.com/ (the list archive)
>>
>> To request this thread, e-mail <[EMAIL PROTECTED]>
>> To unsubscribe, e-mail
>> <[EMAIL PROTECTED]>
>> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
>>
>>
>
>
>
> ---------------------------------------------------------------------
> Before posting, please check:
> http://www.mysql.com/manual.php (the manual)
> http://lists.mysql.com/ (the list archive)
>
> To request this thread, e-mail <[EMAIL PROTECTED]>
> To unsubscribe, e-mail
> <[EMAIL PROTECTED]>
> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
>
>
---------------------------------------------------------------------
Before posting, please check:
http://www.mysql.com/manual.php (the manual)
http://lists.mysql.com/ (the list archive)
To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php