* RAID performance limited by single kernel thread mdX_raid6 ?
@ 2024-04-12 17:35 Tom Crane
2024-04-20 1:59 ` Yu Kuai
0 siblings, 1 reply; 2+ messages in thread
From: Tom Crane @ 2024-04-12 17:35 UTC (permalink / raw
To: linux-raid; +Cc: T.Crane
We have a newly built RAID6 array built using the LVM subsystem. It
comprises 12-off 1TB NVME drives and 12 stripes to match the number of
drives.
The performance seems to be limited by the single [mdX_raid6] kernel
thread which 'top' shows is maxed out at 100% CPU (on a single core)
when performance testing the array.
The kernel is a fairly old one (Redhat's 3.10.0-1160.114.2.el7.x86_64)
on a Centos7 system. A cursory googling of the topic does not suggest
that later kernels (e.g. Redhat's 5.14.0-362.18.1.el9_3.x86_64 should we
upgrade the system to an EL9 distro) have multi-threaded this... Can
anyone comment ?
Many thanks
Tom Crane.
ps. I'm not subscribed to this ML so please CC me when replying.
This email, its contents and any attachments are intended solely for the addressee and may contain confidential information. In certain circumstances, it may also be subject to legal privilege. Any unauthorised use, disclosure, or copying is not permitted. If you have received this email in error, please notify us and immediately and permanently delete it. Any views or opinions expressed in personal emails are solely those of the author and do not necessarily represent those of Royal Holloway, University of London. It is your responsibility to ensure that this email and any attachments are virus free.
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: RAID performance limited by single kernel thread mdX_raid6 ?
2024-04-12 17:35 RAID performance limited by single kernel thread mdX_raid6 ? Tom Crane
@ 2024-04-20 1:59 ` Yu Kuai
0 siblings, 0 replies; 2+ messages in thread
From: Yu Kuai @ 2024-04-20 1:59 UTC (permalink / raw
To: Tom Crane, linux-raid; +Cc: yukuai (C)
Hi,
在 2024/04/13 1:35, Tom Crane 写道:
> We have a newly built RAID6 array built using the LVM subsystem. It
> comprises 12-off 1TB NVME drives and 12 stripes to match the number of
> drives.
>
> The performance seems to be limited by the single [mdX_raid6] kernel
> thread which 'top' shows is maxed out at 100% CPU (on a single core)
> when performance testing the array.
No, IO fast path is not involved with the daemon thread. Daemon thread
is used to handle IO errer, sync IO, update sb, handle spares...
Thanks,
Kuai
>
> The kernel is a fairly old one (Redhat's 3.10.0-1160.114.2.el7.x86_64)
> on a Centos7 system. A cursory googling of the topic does not suggest
> that later kernels (e.g. Redhat's 5.14.0-362.18.1.el9_3.x86_64 should we
> upgrade the system to an EL9 distro) have multi-threaded this... Can
> anyone comment ?
>
>
> Many thanks
>
> Tom Crane.
>
> ps. I'm not subscribed to this ML so please CC me when replying.
>
> This email, its contents and any attachments are intended solely for the
> addressee and may contain confidential information. In certain
> circumstances, it may also be subject to legal privilege. Any
> unauthorised use, disclosure, or copying is not permitted. If you have
> received this email in error, please notify us and immediately and
> permanently delete it. Any views or opinions expressed in personal
> emails are solely those of the author and do not necessarily represent
> those of Royal Holloway, University of London. It is your responsibility
> to ensure that this email and any attachments are virus free.
>
> .
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2024-04-20 1:59 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-04-12 17:35 RAID performance limited by single kernel thread mdX_raid6 ? Tom Crane
2024-04-20 1:59 ` Yu Kuai
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.