zfs discuss - Nov 2009 - Manual drive failure?

So, I''ve done a bit of research and RTFM, and haven''t found an
answer.  If
I''ve missed something obvious, please point me in the right direction.

Is there a way to manually fail a drive via ZFS?  (this is a raid-z2
raidset)  In my case, I''m pre-emptively replacing old drives with
newer,
faster, larger drives.  So far, I''ve only been able to come up with two
solutions to the issue, neither of which is very graceful.

The first option is to simply yank the old drive out of the chassis.  I
could go on at-length about why I dislike doing that, but I think it''s
safe
to say everyone agrees this isn''t a good option.

The second option is to export the zpool, then I can cfgadm  -c disconnect
the drive, and finally gracefully pull it from the system.  Unfortunately,
this means my data has to go offline.  While that''s not a big deal for
a
home box, it is for something in the enterprise with uptime concerns.
>From my experimentation, you can''t disconnect or unconfigure a
drive that is
part of a live zpool.  So, is there a way to tell zfs to pre-emptively fail
it so that you can use cfgadm to put the drive into a state for a graceful
hotswap?  Am I just missing something obvious?  Detach seems to only apply
to mirrors and hot spares.

--Tim
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20091111/6256cffe/attachment.html>

Tim,

I think you''re looking for zpool offline:

     zpool offline [-t] pool device ...

         Takes the specified physical device offline.  While  the
         device  is  offline, no attempt is made to read or write
         to the device.

         This command is not applicable to spares or  cache  dev-
         ices.

         -t    Temporary. Upon  reboot,  the  specified  physical
               device reverts to its previous state.


Ed Plese

On Wed, Nov 11, 2009 at 12:15 PM, Tim Cook <tim at cook.ms>
wrote:
> So, I''ve done a bit of research and RTFM, and haven''t
found an answer.? If
> I''ve missed something obvious, please point me in the right
direction.
>
> Is there a way to manually fail a drive via ZFS?? (this is a raid-z2
> raidset)? In my case, I''m pre-emptively replacing old drives with
newer,
> faster, larger drives.? So far, I''ve only been able to come up
with two
> solutions to the issue, neither of which is very graceful.
>
> The first option is to simply yank the old drive out of the chassis.? I
> could go on at-length about why I dislike doing that, but I think
it''s safe
> to say everyone agrees this isn''t a good option.
>
> The second option is to export the zpool, then I can cfgadm? -c disconnect
> the drive, and finally gracefully pull it from the system.? Unfortunately,
> this means my data has to go offline.? While that''s not a big deal
for a
> home box, it is for something in the enterprise with uptime concerns.
>
> From my experimentation, you can''t disconnect or unconfigure a
drive that is
> part of a live zpool.? So, is there a way to tell zfs to pre-emptively fail
> it so that you can use cfgadm to put the drive into a state for a graceful
> hotswap?? Am I just missing something obvious?? Detach seems to only apply
> to mirrors and hot spares.
>
> --Tim
>
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
>
>