AW: ZFS: Corrupted pool metadata after adding vdev to a pool - no opportunity to rescue data from healthy vdevs? Remove a vdev? Rewrite metadata?

Thomas Göllner (Newsletter) Newsletter at goelli.de
Wed Sep 12 05:47:06 UTC 2012


>Looks pretty bad, but you can try readonly mode with transaction rollback and recheck:
>
>     zpool import -f -R /GoelliZFS1 -N -o readonly=on -F -X GoelliZFS1

Thanks for your suggestion. I think it was a bit like "zpool import -fFXa" so I'm not surprised that it has the same result after ca. 30 seconds:
	cannot import 'GoelliZFS1': one or more devices is currently unavailable

If I can't tell the ZFS to ignore or delete the missing vdev, perhaps I can replace the missing with a "new" one? I have the three 1.5TB discs I wanted to add to the pool. So I can create a raidz-vdev out of them and replace the missing-vdev with the new-raidz-vdev...  Can someone tell me how I am supposed to do this via CLI?

Another idea is this: If I delete the pool, is there a way to take the three 3TB discs with my data and create a "new-old" vdev/pool out of them? I think of this because of the constructed case I mentioned...

I still can't accept, that there's no way to rewrite or edit the metadata ;-)



More information about the freebsd-fs mailing list