[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#598323: linux-image- Servers reboot on heavy load on DRBD+OCFS2 partition

On Fri, Oct 22, 2010 at 10:32:24AM +0400, Proskurin Kirill wrote:
> Hello!
> Sorry for such big delay - I was ill and then on vacation.
> I you still have an interest in this problem - I have new results.
> On 01/10/10 06:49, Ben Hutchings wrote:
>> On Thu, 2010-09-30 at 19:10 +0400, Proskurin Kirill wrote:
>> [...]
>>> Summary:
>>> Kernel: 2.6.36-rc5 SMP x86_64 (from experimental)
>>> DRBD-utils-8.3.8(from experimental)
>>> OCFS2-1.4.4-3(from testing)
>> ocfs2 is already included in the kernel package and you should use that.
> OCFS2-1.4.4-3(from testing) - it is a userspace utility like mkfs.ocfs2.  
> Of course I use driver from kernel.

OK, good.

>>> While update(aptitude safe-upgrade) first node I get kernel panic.
>>> Screenshot in attachment.
>> [...]
>> This panic shows "Tainted: G D" which means there was a previous "oops"
>> message.  You need to record the first one.
> Well I not got it twice.
> I can confirm what on configuration above(all testing + kernel  
> 2.6.36-rc5) I don`t got a reboot. iozone complete successfully without  
> any problems so yes - it is a kernel relaited problem. I retest it on  
> latest 2.6.32 from testing - and got reboot.
> So... what should I do now?

I'm sorry but I don't have any idea where the problem is.  So far as I
can see, there are no bug fixes to drbd or ocfs2 in 2.6.36-rc5 that are
not also in  Maybe the bug is elsewhere and just triggered by
this combination of storage driver and filesystem.  Or, given that you
said that even 2.6.36-rc5 did crash once, it could be that the hardware
is unreliable.

So there are two things you could try, but I am not very hopeful:
1. Run a RAM test such as memtest86+.
2. Use 'git bisect' to find the change that makes the difference.
   Normally you would use this to find when a bug was introduced, but
   you can also use it to find when a bug was fixed if you reverse the
   'good' and 'bad' labels.
   See <http://book.git-scm.com/5_finding_issues_-_git_bisect.html>.



Ben Hutchings
We get into the habit of living before acquiring the habit of thinking.
                                                              - Albert Camus

Attachment: signature.asc
Description: Digital signature

Reply to: