occasional hard lockup on raid on nvme disks (fio processes hung)

occasional hard lockup on raid on nvme disks (fio processes hung)

[Nikhil Kshirsagar]

Hello,

I'm referring to this old email I found in the linux-raid history. Can
anyone comment on if this issue has had further discussion, or
analysis ?

I see subsequent follow ups at spinics.net/lists/raid/msg52994.html and

spinics.net/lists/raid/msg52970.html

spinics.net/lists/raid/msg53461.html

but no further discussion.

I was wondering if this has been reported in RHEL or has any
associated bz's or investigation.

Thanks,
Nikhil.

Re: occasional hard lockup on raid on nvme disks (fio processes hung)

[Artur Paszkiewicz]

On 08/28/2017 08:18 AM, Nikhil Kshirsagar wrote:
> Hello,
> 
> I'm referring to this old email I found in the linux-raid history. Can
> anyone comment on if this issue has had further discussion, or
> analysis ?
> 
> I see subsequent follow ups at spinics.net/lists/raid/msg52994.html and
> 
> spinics.net/lists/raid/msg52970.html
> 
> spinics.net/lists/raid/msg53461.html
> 
> but no further discussion.
> 
> I was wondering if this has been reported in RHEL or has any
> associated bz's or investigation.

It appears that this was recently fixed by:
184a09eb9a2f ("md/raid5: preserve STRIPE_ON_UNPLUG_LIST in break_stripe_batch_list")

I submitted a RHEL BZ for this:
https://bugzilla.redhat.com/show_bug.cgi?id=1497215

Regards,
Artur