ZFS re attaching failed device to pool

Alejandro Imass aimass at yabarana.com
Wed Nov 7 02:54:45 UTC 2018


On Tue, Nov 6, 2018 at 2:14 PM Kevin P. Neal <kpn at neutralgood.org> wrote:

> I do suggest a backup to some other disk that you then take offline.  I
> don't know what will happen if you have disk trouble while a removal is
> in process. That code, if I'm not insane, is very new.
>
>
>
Thank you all for taking the time to respond.

So what is the the correct approach here? Would this suggestion by Kevin
Neal have worked for re-attaching a good disk that had been ousted for a
faulty cable?

If so, what is the procedure for updating the handbook so that other people
that encounter this particular situation can recover more easily?

<quote>
For a simple mirror, if you can handle the risk, you can detach the
unavailable disk and then reattach it. Once the mirror has been dropped
down to a single disk (ie, not a mirror anymore) then you can _attach_ the
disk that was part of the mirror before. Just make sure you "attach" it,
don't use any other verbs.
</quote>

At this point I will just backup, reconstruct the array and restore.

Thanks again!

-- 

Alex


More information about the freebsd-questions mailing list