All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andrew Morton <akpm@linux-foundation.org>
To: Nikola Ciprich <extmaillist@linuxbox.cz>
Cc: linux-kernel@vger.kernel.org, linux-scsi@vger.kernel.org,
	Nick Cheng <nick.cheng@areca.com.tw>,
	Erich Chen <erich@areca.com.tw>
Subject: Re: arcmsr & areca-1660 - strange behaviour under heavy load
Date: Sun, 24 Feb 2008 16:10:34 -0800	[thread overview]
Message-ID: <20080224161034.f494fc7f.akpm@linux-foundation.org> (raw)
In-Reply-To: <alpine.LRH.0.9999.0802231201380.21313@linuxbox.linuxbox.cz>

On Sat, 23 Feb 2008 12:20:12 +0100 (CET) Nikola Ciprich <extmaillist@linuxbox.cz> wrote:

> Hi,
> 
> I've found strange problem either in arcmsr driver, or maybe in 
> areca-1660 card...
> When system on SAS discs RAID connected to areca-1660 card 
> gets under heavy I/O load, it gets unusable after some time. I can 100% reproduce 
> this, although it needs quite speciffic conditions:
> It can be reproduced on 2x quad core machine, RAM has to be limited to 
> ~192MB to cause heavy paging.
> Only thing needed to cause the problem is to start loop doing kernel 
> compilation using make -j 8 - this loads the system heavily, because of 
> lack of memory. After few correct compile runs the system gets into 
> state when all programs including the basic ones (ls, cp, ..) start 
> crashing... dmesg (when it works) doesn't say anything strange...
> After reboot, the system is OK again.
> I have tested it on different motherboards, with different CPUs, RAMs(all 
> were properly tested with memtest), with two different areca cards and 
> different drives. I can't reproduce the problem on same hardware when 
> using different RAID card (ie adaptec). All testing systems were properly 
> cooled..
> I have tried all available areca firmwares, two different distributions 
> (oracle linux, and centos), and kernels ranging from distribution ones, to last GIT snapshot.
> Could somebody please give me some hints on how to hunt this problem?
> Areca support doesn't seem to be very interested in the problem :-(

(cc's added)

Please get the machine into this state of memory exhaustion then take
copies of the output of the following, and send them via reply-to-all to
this email:

- cat /proc/meminfo

- cat /proc/slabinfo

- dmesg -c > /dev/null ; echo m > /proc/sysrq-trigger ; dmesg -c

Thanks.

  reply	other threads:[~2008-02-25  0:10 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-02-23 11:20 arcmsr & areca-1660 - strange behaviour under heavy load Nikola Ciprich
2008-02-25  0:10 ` Andrew Morton [this message]
2008-02-26  9:35   ` Nikola Ciprich
2008-02-26 10:30     ` nickcheng
2008-02-26 10:30       ` nickcheng
2008-02-26 17:43     ` Andrew Morton
2008-02-26 19:29       ` Nikola Ciprich
2008-02-26 21:04         ` Zan Lynx
2008-02-27  1:53           ` nickcheng
2008-02-27  1:53             ` nickcheng

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20080224161034.f494fc7f.akpm@linux-foundation.org \
    --to=akpm@linux-foundation.org \
    --cc=erich@areca.com.tw \
    --cc=extmaillist@linuxbox.cz \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=nick.cheng@areca.com.tw \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.