Kernel panic during resync

Kernel panic during resync

[Paul Waldo]

Hi all,

I've got a machine with a RAID6 array which hung on me yesterday.  Upon
reboot, mdadm refused to start the array, since it was degraded and
dirty.  The array had 7 drives, and one had previously gone bad.  I'm
running Fedora Core 5.

I rebooted, using the "md-mod.start_dirty_degraded=1" parameter.
Everything seemed to be OK--the sync process started and I watched it go
for about half an hour.  I came back later, and found the machine had a
kernel panic with the message "Kernel panic - not syncing: Fatal
exception in interrupt".  Just to make sure, I restarted and tried to
sync again, but got the same message.

What might be causing this?  Is there any way to recover?  Thanks in
advance!

Paul

Re: Kernel panic during resync

[Neil Brown]

On Friday August 11, pwaldo@waldoware.com wrote:
> Hi all,
> 
> I've got a machine with a RAID6 array which hung on me yesterday.  Upon
> reboot, mdadm refused to start the array, since it was degraded and
> dirty.  The array had 7 drives, and one had previously gone bad.  I'm
> running Fedora Core 5.
> 
> I rebooted, using the "md-mod.start_dirty_degraded=1" parameter.
> Everything seemed to be OK--the sync process started and I watched it go
> for about half an hour.  I came back later, and found the machine had a
> kernel panic with the message "Kernel panic - not syncing: Fatal
> exception in interrupt".  Just to make sure, I restarted and tried to
> sync again, but got the same message.
> 
> What might be causing this?  Is there any way to recover?  Thanks in
> advance!

It is very hard to say without more details.
Was there any stack trace?  Any other possibly related messages that
you can report?

And what kernel (exactly) are you using?

NeilBrown
 

Re: Kernel panic during resync

[Paul Waldo]

Neil Brown wrote:
> On Friday August 11, pwaldo@waldoware.com wrote:
>> Hi all,
>>
>> I've got a machine with a RAID6 array which hung on me yesterday.  Upon
>> reboot, mdadm refused to start the array, since it was degraded and
>> dirty.  The array had 7 drives, and one had previously gone bad.  I'm
>> running Fedora Core 5.
>>
>> I rebooted, using the "md-mod.start_dirty_degraded=1" parameter.
>> Everything seemed to be OK--the sync process started and I watched it go
>> for about half an hour.  I came back later, and found the machine had a
>> kernel panic with the message "Kernel panic - not syncing: Fatal
>> exception in interrupt".  Just to make sure, I restarted and tried to
>> sync again, but got the same message.
>>
>> What might be causing this?  Is there any way to recover?  Thanks in
>> advance!
> 
> It is very hard to say without more details.
> Was there any stack trace?  Any other possibly related messages that
> you can report?
> 
> And what kernel (exactly) are you using?
> 
> NeilBrown

Hi Niel,

Thanks for the reply.  I resolved the issue by using a different disk. 
I had a suspicion that the disk might have some problems, but I figured 
that the sync process would work around it and give at least a warning 
or give me a message saying that the disk was unusable.  I wasn't 
expecting a panic :-O

My recollection of the stack trace was the problem was in a 
raid6-buffer-flush type of place.  Sorry I don't have any more info for 
you...

Paul