zfs discuss - Sep 2011 - All drives intact but vdev UNAVAIL in raidz1

Hi all,

I''m running FreeNAS 7 (based on FreeBSD 7.2) and using ZFS v.13. I had
two 2TB drives and one 1TB drive in a raidz1. I recently bought another 2TB
drive to replace the 1TB drive. Unfortunately, I ran into some serious issues
which are described in the steps that I did below:

First, I scrubbed the disks and found no errors. Next I turned off the computer,
unplugged the 1TB drive and plugged in the new 2TB drive into the same SATA
port. I turned on the machine, and executed ''zpool status''.
The system hung and ignored all interrupt and kill signals, and terminated any
shutdown sequence, so I hard-shutdown the machine. Next I unplugged the new 2TB
drive and plugged in the old 1TB drive. When I enter ''zpool
status'' I see that the pool is UNAVAIL and the vdev is CORRUPTED, and
all three drives are online. I tried to export & import the pool, but was
unable to import as there ''are not enough redundant copies of the
data.''

It seems quite likely that all of the data is intact, and that something
different is preventing me from accessing the pool. What can I do to recover the
pool? I have downloaded the Solaris 11 express livecd if that would be of any
use.

Thanks in advance,
Tyler

On Tue, 6 Sep 2011, Tyler Benster wrote:
> It seems quite likely that all of the data is intact, and that something 
> different is preventing me from accessing the pool. What can I do to 
> recover the pool? I have downloaded the Solaris 11 express livecd if 
> that would be of any use.
Try running zdb -l on the disk and see if the labels are still there. 
Also, could you show us the output of ''zpools status''?
Normally zfs would
not hang if one disk of a raidz group is missing, but it might do that if 
one toplevel is missing.

If the zdb command shows all four labels to be correct, then you can try a 
zpool scrub and see if that resilvers the data for you.