Bug#568922: mdadm: data-check of RAID array md0 results in soft lockup of CPU's

[Jelle de Jong]

Hi Tim,

Tim Small wrote, on 03/04/10 17:01:
> What happens if you disable NCQ?  Seen similar crap on another system
> with WD drives, and I suspect a firmware bug.
> 
> That system was OK with a queue depth of 2 or less.  You can confirm
> this with hdparm -t, I think.
> 
> This problem should really be assigned to the kernel I suppose, as all
> that mdadm does is to tell the kernel to execute the check....

I attached the hdparm information. If I see the problem again I will
try to disable the NCQ of the drives if activated.

Thanks in advance,

Best regards,

Jelle
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: hdparm-info.txt
URL: <http://lists.alioth.debian.org/pipermail/pkg-mdadm-devel/attachments/20100304/a14c72f8/attachment.txt>