1

We had 2 failed drives in our ZFS raidz1 pool. One had been resilvered to a spare and another was just out. The pool was in degraded mode.

I replaced both drives and began the resilver on the non-spare drive. Without thinking, I forced a resilver on the one that was currently running on the spare. The pool is now reporting LOTS of data errors.

Is there any way to recover from this or am I totally screwed? I would think that since the spare is still available, we're still only down 1 drive and the data should technically all be there. Also, did I cause this or did I just uncover existing data corruption?

Derek
  • 165
  • 1
  • 7
  • What **exactly** did you do? You can't force ZFS to "resilver", it does it on it's own. – Chris S Sep 26 '12 at 18:27
  • Obligatory "RAID is not backup" comment. – David Schwartz Sep 26 '12 at 18:42
  • Chris: I did a zpool replace -f app-pool (disk ID) for both disks. – Derek Sep 26 '12 at 18:50
  • Yeah, you're screwed. Can I ask why you would have included the `-f` option? It's one of those things they shouldn't have included in the first place, and there's no valid reason to use it ever. – Chris S Sep 26 '12 at 19:03
  • I did it to the 1st disk because it HAD been part of another pool and it gave me an error. On the 2nd disk it was a mistake. – Derek Sep 26 '12 at 19:24
  • Also David, I've been fighting for a proper backup for months. I work at a government installation. Nothing is easy. – Derek Sep 26 '12 at 19:25
  • I guess I don't understand why forcing a replacement on two disks that were already offline would cause this problem though. It's not like I removed active disks that were being read from. – Derek Sep 27 '12 at 13:20

0 Answers0