SUMMARY: zfs rootpool degraded and migration

I didn't get a single hit from this mail. :-(

The zfs root file system finally reached a point of no return.  I didn't 
have a chance to try reseating the drives or anything.  The machine 
crashed and wouldn't boot again, so we just replaced both disks.

For the migration issue...  I had installed the OS on machine2, then we 
popped it (a single disk at this point) into machine1.  As before, it 
failed to boot, saying that the zfs pool 'rootpool' was last mounted on 
a different machine (machine2).  So, I booted off the network into 
single user.  Then ran 'zpool import -f rootpool'.  Then I rebooted off 
the disk and all was fine.

There may be a better procedure, but this does work.

FYI

Tom Lieuallen
Oregon State University

Tom Lieuallen wrote:
> I have a V120 running Solaris 10 U6.  I decided to use a zfs root file 
> system so that I could mirror without all the solstice pieces.  I 
> installed the OS, then attached the second drive after the OS was up 
> (zpool attach...).  It appears that the resilvering had problems.
> 
> # zpool status
>    pool: rootpool
>   state: DEGRADED
> status: One or more devices has experienced an error resulting in data
>          corruption.  Applications may be affected.
> action: Restore the file in question if possible.  Otherwise restore the
>          entire pool from backup.
>     see: http://www.sun.com/msg/ZFS-8000-8A
>   scrub: resilver completed after 0h41m with 51926 errors on Wed Dec 31 
> 13:05:02
> 2008
> config:
> 
>          NAME          STATE     READ WRITE CKSUM
>          rootpool      DEGRADED    10     0  108K
>            mirror      DEGRADED    10     0  108K
>              c0t0d0s0  FAULTED     20     0     0  too many errors
>              c0t1d0s0  DEGRADED     0     1  216K  too many errors
> 
> errors: Permanent errors have been detected in the following files:
> 
>          /a1/solaris10/include/curl/curl.h
>          /a1/solaris10/include/curl/curlver.h
> ...
> <ad nausea>
> ...
> 
> # zfs list
> NAME                            USED  AVAIL  REFER  MOUNTPOINT
> rootpool                       13.3G  19.9G    94K  /rootpool
> rootpool/ROOT                  4.25G  19.9G    18K  legacy
> rootpool/ROOT/solaris10_6      4.25G  19.9G  4.17G  /
> rootpool/ROOT/solaris10_6/var  83.3M  19.9G  83.3M  /var
> rootpool/a1                    2.75G  5.25G  2.75G  /a1
> rootpool/dump                  2.00G  19.9G  2.00G  -
> rootpool/private                341M  5.67G   341M  /private
> rootpool/swap                     4G  23.9G    16K  -
> 
> # iostat -e
>             ---- errors ---
> device    s/w h/w trn tot
> ramdisk1    0   0   0   0
> sd0         0 129 487 616
> sd3         0 256 877 1133
> nfs1        0   0   0   0
> 
> The system logs report scsi bus resets and read and write errors for 
> both disks.  I assume one of the disks is causing problems for 
> everything on the bus.
> 
> The 'faulted' disk is the one with the original OS.  The files it is 
> reporting problems with are all on rootpool/a1.  There's only ~50,000 of 
> them.  That's just a copy of our /usr/local, so there's nothing there I 
> need.
> 
> Should we just shut the machine down/off, reseat both disks, and hope 
> for the best?  I'm concerned it won't boot again. :-(  Or, if we try to 
> hot-plug one of the disks, it may panic too.
> 
> I have another related question...
> 
> I'd like to prepare another disk or two to slap into this host in case 
> it's not repairable or in case something I do makes things worse. :-) 
> What is the proper procedure for installing a machine with zfs root, 
> then moving that disk to another host?  I did this once before and it 
> complained about the zpool being last used by a different host.  I ended 
> up booting off the net and forcing the import.  To simplify matters, 
> should I boot the temporary install host off the net and export the 
> zpool?  Or should I just be prepared to force the import?
> 
> thank you
> 
> Tom Lieuallen
> Oregon State University
> _______________________________________________
> sunmanagers mailing list
> sunmanagers@sunmanagers.org
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers