4

I have moved a ZFS pool from one SunFire server running some pre-release of Solaris 11 to another server running latest Solaris 10 and now can't import it. It was properly zfs export from the old one. At the same time I migrated one another zpool (extpool2) and it imported just fine. Here is what I get on the new server:

root@kria ~ # zpool import
  pool: extpool
    id: 8401594403030600070
 state: ONLINE
status: The pool is formatted using an older on-disk version.
action: The pool can be imported using its name or numeric identifier, though
        some features will not be available without an explicit 'zpool upgrade'.
config:

        extpool      ONLINE
          raidz1-0   ONLINE
            c2t8d0   ONLINE
            c2t10d0  ONLINE
            c2t12d0  ONLINE
            c2t14d0  ONLINE

root@kria ~ # zpool import -f extpool
cannot import 'extpool': one or more devices is currently unavailable

But they are available - all 4 are online, /dev/ entries are there and they show up in format device list. The extpool zfs version is 31 (I checked that before moving it) and the host supports at least ver 32 (checked its rpool version).

I also did zdb -l and got this failure:

root@kria ~ # zdb -l /dev/dsk/c2t8d0
------------------------------------------
LABEL 0
------------------------------------------
------------------------------------------
LABEL 1
------------------------------------------
failed to unpack label 1
------------------------------------------
LABEL 2
------------------------------------------
------------------------------------------
LABEL 3 - CONFIG MATCHES LABEL 2
------------------------------------------

Any idea what to do?

  • seems like corruption on one of the disks. Have you tried to re-import in the original server? – BitsOfNix Jul 5 '13 at 6:27
  • I would do what @AlexandreAlves is suggesting. First so you can see if the older disk version import is complaining about shows up on the origin server. And secondly if you can import the pool on the origin server, use offline to offline the suspect disk and export the pool again. You could use a sparse file "placeholder" to replace the defective disk (if any). – Tim Jul 5 '13 at 20:40
  • 1
    Thanks for the suggestion. I was able to mount the array in another Solaris 11 box. That made me suspect that there is some incompatibility between S11 and S10 - turned out to be the zfs encryption. Although it's supported from ZFS version 31 on Solaris 11 it isn't supported on Solaris 10 at all even though it supports up to version 32 :( – MLu Jul 5 '13 at 23:51
3

Apparently there is incompatibility between Solaris 11 and Solaris 10 - ZFS Encryption and ZFS Deduplication. Although they are supported from ZFS version 31 and 21 respectively on Solaris 11 they aren't supported on Solaris 10 at all even though it supports up to version 32 :(

Hint: zpool upgrade -v for a list of supported features.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.