redhat-list Digest, Vol 83, Issue 2

[John Dalbec]

I got Alt+SysRq+t output, but it looks corrupted in /var/log/messages. 
I suspect that syslogd couldn't keep up with klogd and the ring buffer 
wrapped.  The system is not starved for CPU, but then I have 24 cores. 
If 32-bit + > 16GB is trouble then why does the kernel-hugemem package 
even exist?  Or is that actually a 64-bit kernel?
Thanks,
John

> Date: Mon, 3 Jan 2011 18:21:17 +0100
> From: (Imed Chihi) ???? ?????? 	<imed.chihi at gmail.com>
> To: redhat-list at redhat.com
> Subject: Re: RHEL4 Sun Java Messaging Server deadlock (was:
> 	redhat-list	Digest, Vol 83, Issue 1)
> Message-ID:
> 	<AANLkTi=Xo8n8YeM6VQuv3uDVrG0HghA0UDftsuKCPpDR at mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
>
>> Date: Mon, 03 Jan 2011 10:51:58 -0500
>> From: John Dalbec<jpdalbec at ysu.edu>
>> To: redhat-list at redhat.com
>> Subject: RHEL4 Sun Java Messaging Server deadlock
>> Message-ID:<4D21F09E.3070100 at ysu.edu>
>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>>
>> Sun Java(tm) System Messaging Server 6.2-9.20 (built Jul 15 2010)
>> libimta.so 6.2-9.20 (built 01:27:24, Jul 15 2010)
>> Linux myysumail.ysu.edu 2.6.9-89.31.1.ELhugemem #1 SMP Mon Oct 4
>> 22:04:11 EDT 2010 i686 i686 i386 GNU/Linux
>>
>> The IMAPd process appears to get into a deadlock with pdflush and
>> kjournald. ?The mailboxes were initially stored on a disk partition but
>> I'm in the process of migrating them to a LVM logical partition. ?Both
>> partitions are ext3. ?I worry about using ext2 because a fsck takes
>> about 90 minutes.
>
> Try collecting the output of Alt+SysRq+m and Alt+SysRq+t when this
> deadlock occurs.  I remotely recall a case where one of the VM threads
> starves the system of CPU cycles on RHEL 4.
>
>> The server has 32GB RAM. ?Right now Committed_AS is around 4GB, but the
>> system is lightly loaded. ?I'm running a 32-bit kernel because the
>> application vendor doesn't support 64-bit.
>
> Running a 32-bit kernel with more than, say, 16 GB of RAM will always
> get you into trouble.  RHEL 4 does have multilib support, so you
> should be able to run a 32-bit application atop a 64-bit installation.
>
>   -Imed
>