ZFS unable to import pool

[Johan Hendriks]

op 23-04-14 14:00, Hugo Lombard schreef:
> On Wed, Apr 23, 2014 at 12:18:37PM +0200, Johan Hendriks wrote:
>> Did you in the past add an extra disk to the pool?
>> This could explain the whole issue as the pool is missing a whole vdev.
>>
> I agree that there's a vdev missing...
>
> I was able to "simulate" the current problematic import state (sans
> failed "disk7", since that doesn't seem to be the stumbling block) by
> adding 5 disks [1] to get to here:
>
>    # zpool status test
>      pool: test
>     state: ONLINE
>      scan: none requested
>    config:
>    
> 	  NAME        STATE     READ WRITE CKSUM
> 	  test        ONLINE       0     0     0
> 	    raidz1-0  ONLINE       0     0     0
> 	      md3     ONLINE       0     0     0
> 	      md4     ONLINE       0     0     0
> 	      md5     ONLINE       0     0     0
> 	      md6     ONLINE       0     0     0
> 	      md7     ONLINE       0     0     0
> 	    raidz1-2  ONLINE       0     0     0
> 	      md8     ONLINE       0     0     0
> 	      md9     ONLINE       0     0     0
> 	      md10    ONLINE       0     0     0
> 	      md11    ONLINE       0     0     0
> 	      md12    ONLINE       0     0     0
> 	  logs
> 	    md1s1     ONLINE       0     0     0
> 	  cache
> 	    md1s2     ONLINE       0     0     0
>    
>    errors: No known data errors
>    #
>
> Then exporting it, and removing md8-md12, which results in:
>
>    # zpool import
>       pool: test
>         id: 8932371712846778254
>      state: UNAVAIL
>     status: One or more devices are missing from the system.
>     action: The pool cannot be imported. Attach the missing
> 	  devices and try again.
>       see: http://illumos.org/msg/ZFS-8000-6X
>     config:
>    
> 	  test         UNAVAIL  missing device
> 	    raidz1-0   ONLINE
> 	      md3      ONLINE
> 	      md4      ONLINE
> 	      md5      ONLINE
> 	      md6      ONLINE
> 	      md7      ONLINE
> 	  cache
> 	    md1s2
> 	  logs
> 	    md1s1      ONLINE
>    
> 	  Additional devices are known to be part of this pool, though their
> 	  exact configuration cannot be determined.
>    #
>
> One more data point:  In the 'zdb -l' output on the log device it shows
>
>    vdev_children: 2
>
> for the pool consisting of raidz1 + log + cache, but it shows
>
>    vdev_children: 3
>
> for the pool with raidz1 + raidz1 + log + cache.  The pool in the
> problem report also shows 'vdev_children: 3' [2]
>
>
>
> [1] Trying to add a single device resulted in zpool add complaining
> with:
>
>    mismatched replication level: pool uses raidz and new vdev is disk
>
> and trying it with three disks said:
>
>    mismatched replication level: pool uses 5-way raidz and new vdev uses 3-way raidz
>
>
> [2] http://lists.freebsd.org/pipermail/freebsd-fs/2014-April/019340.html
>
But you can force it....
If you force it, it will add a vdev not the same as the current vdev. So 
you will have a raidz1 and a single no parity vdev in the pool. If you 
destroy the single disk vdev then you will get a pool which can not be 
repaired as far as I know.

regards
Johan