* VERY slow mdadm recovery speed 12KB/s
[not found] <a889ac7d1003042123x7473976dw8ffe2af6f2844afc@mail.gmail.com>
@ 2010-03-05 5:27 ` Benjamin Kingston
2010-03-05 8:55 ` Michael Evans
0 siblings, 1 reply; 2+ messages in thread
From: Benjamin Kingston @ 2010-03-05 5:27 UTC (permalink / raw)
To: linux-raid
I did some musical HDDs to get my "best" drives to hold a critical md
device, then one failed after a power surge. So I bought a new disk
and added it, then accidentally marked it as failed in the middle of
recovery. I then removed the disk and added it again. I got the
message "/dev/sdd re-added to /dev/md1"
When I do a cat /proc/mdstat the recovery speed is at 55KB/s and
subsequent views of /proc/mdstat show that the speed goes all the way
to 12KB/s or lower. I've since removed and re-added the drive a few
times, but the same thing happens every time and if I leave everything
alone while it resyncs, my CPU usage sky rockets to something like
16.8 23.4 40.5.
Any help would be really appreciated as I don't want to loose the
"good" drive. I know both have good speeds, because I can write at
4MB/s to ext4/luks on md1 and 15MB/s on /dev/sdd with dd.
P.S. could someone also explain to me the difference between md0 and md_d0?
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: VERY slow mdadm recovery speed 12KB/s
2010-03-05 5:27 ` VERY slow mdadm recovery speed 12KB/s Benjamin Kingston
@ 2010-03-05 8:55 ` Michael Evans
0 siblings, 0 replies; 2+ messages in thread
From: Michael Evans @ 2010-03-05 8:55 UTC (permalink / raw)
To: Benjamin Kingston; +Cc: linux-raid
md_d0 is a side-effect of trying to auto-run a conflictingly numbered
device, or non-native host (from a computer with a different hostname)
array. Otherwise the device is exactly as normal. If you enumerate
your arrays, even only by UUID and the device name/number you want in
/etc/mdadm.conf it will be obeyed.
Unfortunately I have no clue why your recovery is going so slowly.
Please post:
cat /etc/mdadm.conf
mdadm -Esvv
Possibly also the contents of
dmesg
if it seems even remotely relevant, and also:
for ii in /sys/block/md*/md/sync_speed* ; do \
echo -n "$ii " ; cat $ii ; done
On Thu, Mar 4, 2010 at 9:27 PM, Benjamin Kingston <ben@bencloud.net> wrote:
> I did some musical HDDs to get my "best" drives to hold a critical md
> device, then one failed after a power surge. So I bought a new disk
> and added it, then accidentally marked it as failed in the middle of
> recovery. I then removed the disk and added it again. I got the
> message "/dev/sdd re-added to /dev/md1"
>
> When I do a cat /proc/mdstat the recovery speed is at 55KB/s and
> subsequent views of /proc/mdstat show that the speed goes all the way
> to 12KB/s or lower. I've since removed and re-added the drive a few
> times, but the same thing happens every time and if I leave everything
> alone while it resyncs, my CPU usage sky rockets to something like
> 16.8 23.4 40.5.
>
> Any help would be really appreciated as I don't want to loose the
> "good" drive. I know both have good speeds, because I can write at
> 4MB/s to ext4/luks on md1 and 15MB/s on /dev/sdd with dd.
>
> P.S. could someone also explain to me the difference between md0 and md_d0?
> --
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2010-03-05 8:55 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <a889ac7d1003042123x7473976dw8ffe2af6f2844afc@mail.gmail.com>
2010-03-05 5:27 ` VERY slow mdadm recovery speed 12KB/s Benjamin Kingston
2010-03-05 8:55 ` Michael Evans
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).