On Tue, Apr 26, 2011 at 4:59 PM, Richard Elling
<richard.ell...@gmail.com> wrote:
>
> On Apr 26, 2011, at 8:22 AM, Cindy Swearingen wrote:
>
>> Hi--
>>
>> I don't know why the spare isn't kicking in automatically, it should.
>
> This can happen if the FMA agents aren't working properly.
>
> FYI, in NexentaStor we have added a zfs-monitor FMA agent to check the
> health of disks in use for ZFS and notice when they are no longer responding
> to reads.
I just recently (this past week) had a very similar failure. zpool
consisting of two raidz2 vdevs and two hot spare drives. Each raidz2
vdev consists of 10 drives (I know, not the best layout, but the
activity is large sequential writes and reads and we needed the
capacity). We had a drive fail in one of the vdevs and one of the hot
spares automatically went into action (the special spare device within
the vdev came into being and the hot spare drive resilvered). A short
time later a second drive in the same vdev failed. No action by any
hot spare. The system was running Solaris 10U8 with no additional
patches.
I opened a case with Oracle and they told me that the hot spare
*should* have dealt with the second failure. We replaced the first
(hot spared) drive with zpool replace and it resilvered fine. Then we
replaced the second (non hot spared) drive with zpool replace and the
system hung. I suspected the mpt (multipathing) driver for the SATA
drives in the J4400, there have been some huge improvements in that
driver since 10U8. After rebooting the drive appeared replaced and was
resilvering.
Oracle support chocked the hot spare issue up to an FMA problem
but could not duplicate it in the lab. We have since upgraded to 10U9
+ the latest CPU (April 2011) and are hoping both the hot spare issue
and the mpt drive issue are fixed.
--
{--------1---------2---------3---------4---------5---------6---------7---------}
Paul Kraus
-> Senior Systems Architect, Garnet River ( http://www.garnetriver.com/ )
-> Sound Coordinator, Schenectady Light Opera Company (
http://www.sloctheater.org/ )
-> Technical Advisor, RPI Players
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
