ZFS from FreeBSD -> Indiana -> FreeBSD - some problem
Adam McDougall
mcdouga9 at egr.msu.edu
Wed Nov 14 17:52:20 PST 2007
Henri Hennebert wrote:
> Henri Hennebert wrote:
>> Hello,
>>
>> As seen in another thread, I was curious about a zpool created under
>> FreeBSD being available under opensolaris.
>>
>> So I boot indiana and do:
>>
>> zpool import -f -o ro pool2
>>
>> zpool scrub pool2
>>
>> all ended properly...
>>
>> shutdown.
>>
>> When I reboot with FreeBSD (7.0-BETA2) I get:
>>
>> zpool status pool2
>> pool: pool2
>> state: UNAVAIL
>> status: One or more devices could not be opened. There are insufficient
>> replicas for the pool to continue functioning.
>> action: Attach the missing device and online it using 'zpool online'.
>> see: http://www.sun.com/msg/ZFS-8000-D3
>> scrub: none requested
>> config:
>>
>> NAME STATE READ WRITE CKSUM
>> pool2 UNAVAIL 0 0 0 insufficient
>> replicas
>> raidz1 UNAVAIL 0 0 0 insufficient
>> replicas
>> dsk/c10d0p3 UNAVAIL 0 0 0 cannot open
>> dsk/c11d0p3 UNAVAIL 0 0 0 cannot open
>>
>> Is it possible to replace dsk/c10d0p3 by ad4s3 and dsk/c11d0p3 by
>> ad6s3 as it must be under FreeBSD or does I need to make some entry
>> in /etc/devfs.rules to create dsk/c10d0p3 and dsk/c11d0p3 (if at all
>> possible?)
>>
>> Isn't it a interesting challenge ?
> I got it:
>
> zpool export pool2 && zpool import -f pool2
>
> do the trick
>
> sorry for the noise :-|
>
> Henri
>
For what its worth, I ran across a similar issue. I moved a scsi card
in my server which caused da2 and da3 (each with a separate zfs pool
with no redundancy) become swapped, which I did not predict. ZFS was
completely confused by this, and rather than swap cables and reboot, I
decided to try the export and import. Worked fine. I was a little
dissapointed and surprised that I had to take action, but more surprised
that such a simple fix was possible (yet predicted) when the error
message on sun's website basically indicated 'you are totally screwed,
game over, data lost'. I wasn't in a situation were I would have lost
any valuable data. Maybe it was terminally confused because both
devices it wanted were in use by the other 'failed' pool.
More information about the freebsd-current
mailing list