Problems replacing failing drive in ZFS pool

[Dan Langille]

On 7/21/2010 2:54 AM, Charles Sprickman wrote:
> On Wed, 21 Jul 2010, Charles Sprickman wrote:
>
>> On Tue, 20 Jul 2010, alan bryan wrote:
>>
>>>
>>>
>>> --- On Mon, 7/19/10, Dan Langille <dan at langille.org> wrote:
>>>
>>>> From: Dan Langille <dan at langille.org>
>>>> Subject: Re: Problems replacing failing drive in ZFS pool
>>>> To: "Freddie Cash" <fjwcash at gmail.com>
>>>> Cc: "freebsd-stable" <freebsd-stable at freebsd.org>
>>>> Date: Monday, July 19, 2010, 7:07 PM
>>>> On 7/19/2010 12:15 PM, Freddie Cash
>>>> wrote:
>>>> > On Mon, Jul 19, 2010 at 8:56 AM, Garrett
>>>> Moore<garrettmoore at gmail.com> wrote:
>>>> >> So you think it's because when I switch from the
>>>> old disk to the new disk,
>>>> >> ZFS doesn't realize the disk has changed, and
>>>> thinks the data is just
>>>> >> corrupt now? Even if that happens, shouldn't the
>>>> pool still be available,
>>>> >> since it's RAIDZ1 and only one disk has gone
>>>> away?
>>>> > > I think it's because you pull the old drive, boot with
>>>> the new drive,
>>>> > the controller re-numbers all the devices (ie da3 is
>>>> now da2, da2 is
>>>> > now da1, da1 is now da0, da0 is now da6, etc), and ZFS
>>>> thinks that all
>>>> > the drives have changed, thus corrupting the
>>>> pool.  I've had this
>>>> > happen on our storage servers a couple of times before
>>>> I started using
>>>> > glabel(8) on all our drives (dead drive on RAID
>>>> controller, remove
>>>> > drive, reboot for whatever reason, all device nodes
>>>> are renumbered,
>>>> > everything goes kablooey).
>>>>
>>>> Can you explain a bit about how you use glabel(8) in
>>>> conjunction with ZFS?  If I can retrofit this into an
>>>> exist ZFS array to make things easier in the future...
>>>>
>>>> 8.0-STABLE #0: Fri Mar  5 00:46:11 EST 2010
>>>>
>>>> ]# zpool status
>>>>   pool: storage
>>>> state: ONLINE
>>>> scrub: none requested
>>>> config:
>>>>
>>>>         NAME STATE     READ WRITE CKSUM
>>>>         storage
>>>>    ONLINE
>>>>    0     0
>>>>    0
>>>>           raidz1 ONLINE       0
>>>>    0     0
>>>>             ad8
>>>>    ONLINE
>>>>    0     0
>>>>    0
>>>>             ad10 ONLINE       0
>>>>    0     0
>>>>             ad12 ONLINE       0
>>>>    0     0
>>>>             ad14 ONLINE       0
>>>>    0     0
>>>>             ad16 ONLINE       0
>>>>    0     0
>>>>
>>>> > Of course, always have good backups.  ;)
>>>>
>>>> In my case, this ZFS array is the backup.  ;)
>>>>
>>>> But I'm setting up a tape library, real soon now....
>>>>
>>>> -- Dan Langille - http://langille.org/
>>>> _______________________________________________
>>>> freebsd-stable at freebsd.org
>>>> mailing list
>>>> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
>>>> To unsubscribe, send any mail to
>>>> "freebsd-stable-unsubscribe at freebsd.org"
>>>>
>>>
>>> Dan,
>>>
>>> Here's how to do it after the fact:
>>>
>>> http://unix.derkeiler.com/Mailing-Lists/FreeBSD/current/2009-07/msg00623.html
>>>
>>
>> Two things:
>>
>> -What's the preferred labelling method for disks that will be used
>> with zfs these days? geom_label or gpt labels? I've been using the
>> latter and I find them a little simpler.
>>
>> -I think that if you already are using gpt partitioning, you can add a
>> gpt label after the fact (ie: gpart -i index# -l your_label adaX).
>> "gpart list" will give you a list of index numbers.
>
> Oops.
>
> That should be "gpart modify -i index# -l your_label adax".

I'm not using gpt partitioning.  I think I'd like to try that.  To do 
just that, I've ordered two more HDD.  They'll be arriving today.

-- 
Dan Langille - http://langille.org/