degraded zfs slowdown

[Dan Naumov]

On Tue, Mar 16, 2010 at 10:23 AM, Randy Bush <randy at psg.com> wrote:
>>> as the ghostly bad drive does not have a unique name, [how] can i tell
>>> zfs to remove it for the moment?
>> Have you tried running a scrub on the pool and THEN trying to offline
>> and detach (zpool commands, not physical actions) the ghost drive?
>
>    psg.com:/root# zpool scrub tank
>    psg.com:/root# zpool status
>      pool: tank
>     state: DEGRADED
>    status: One or more devices could not be used because the label is missing or
>            invalid.  Sufficient replicas exist for the pool to continue
>            functioning in a degraded state.
>    action: Replace the device using 'zpool replace'.
>       see: http://www.sun.com/msg/ZFS-8000-4J
>     scrub: scrub in progress for 0h0m, 0.00% done, 169h12m to go
>    config:
>
>            NAME        STATE     READ WRITE CKSUM
>            tank        DEGRADED     0     0     0
>              mirror    DEGRADED     0     0     0
>                twed1   FAULTED      0     3     0  corrupted data
>                twed1   ONLINE       0     0     0
>
>    errors: No known data errors
>
> still do not know how to detach bad drive as it does not have a unique
> name

As you can see from your own message, the scrub hasn't finished, it
has just started. What happens after the scrub is finished? If both
twed1's are still visible at this point (after a finished scrub), what
happens if you attempt to "zfs detach tank twed1" the ghost drive, do
you get an error or does it recognize and correctly remove the ghost
twed1?

- Sincerely,
Dan Naumov