Help with Raidz2 array?

[monogoat]

So I've been running a 7 drive Raidz2 array under Ubuntu as a media server, and one morning I could no longer access the files. A zpool status told me that the array failed because of a bad intent log, one drive listed as UNAVAIL due to corrupted data. In my searching of various forum posts around the interwebs, I tried a zfs clear, but it failed. I then did a zfs export which appears to have worked fine, but zfs import will not work. Here is my current situation:

root@LinuxServer:~# zpool import -m
pool: fileserver
id: 7093643436194074128
state: ONLINE
status: One or more devices contains corrupted data.
action: The pool can be imported using its name or numeric identifier.
see: http://zfsonlinux.org/msg/ZFS-8000-4J
config:

fileserver ONLINE
raidz2-0 ONLINE
ata-ST4000DM000-1F2168_W300D13P UNAVAIL corrupted data
ata-ST4000DM000-1F2168_W300PPAV ONLINE
ata-ST4000DM000-1F2168_W300TNP6 ONLINE
ata-ST4000DX000-1CL160_Z1Z02WWY ONLINE
ata-ST4000DX000-1CL160_Z1Z03D6X ONLINE
ata-ST4000DX000-1CL160_Z1Z0491E ONLINE
ata-ST4000DX000-1CL160_Z1Z05NX1 ONLINE
root@LinuxServer:~# zpool import -m fileserver
cannot import 'fileserver': pool may be in use from other system
use '-f' to import anyway
root@LinuxServer:~# zpool import -m -f fileserver
cannot import 'fileserver': one or more devices is currently unavailable


If I pull the offending drive from the server, it drops the corrupted data part of the status, but gives the same issues on import. Anyone have any input on this situation? I'd really prefer not to have to recreate everything that was on the server. Downloading 200gbs of stuff from crashplan isn't a problem, but ripping 8tb of movies takes some time. Thanks for your time

[substr]

Have you tried zpool import -f -F fileserver

Might also be zpool import -fF fileserver

This is supposed to try and throw out the last txg, but not sure if that also works around a problem with the ZIL.

Also recommend you run a memory test, maybe first thing, just in case you've got some bad hardware. I think Ubuntu includes a memtest on the boot/install media.

[monogoat]

I will run a memtest on the system. This is what your command gives:

root@LinuxServer:~# zpool import -m
pool: fileserver
id: 7093643436194074128
state: ONLINE
status: One or more devices contains corrupted data.
action: The pool can be imported using its name or numeric identifier.
see: http://zfsonlinux.org/msg/ZFS-8000-4J
config:

fileserver ONLINE
raidz2-0 ONLINE
ata-ST4000DM000-1F2168_W300D13P UNAVAIL corrupted data
ata-ST4000DM000-1F2168_W300PPAV ONLINE
ata-ST4000DM000-1F2168_W300TNP6 ONLINE
ata-ST4000DX000-1CL160_Z1Z02WWY ONLINE
ata-ST4000DX000-1CL160_Z1Z03D6X ONLINE
ata-ST4000DX000-1CL160_Z1Z0491E ONLINE
ata-ST4000DX000-1CL160_Z1Z05NX1 ONLINE
root@LinuxServer:~# zpool import -f -F fileserver
cannot import 'fileserver': I/O error
Destroy and re-create the pool from
a backup source.


This kinda sucks, for some reason I am a magnet for problems. I had hoped that having a 2 drive fault tolerance would be good, and here I have an issue without even having one drive die.

[crowi]

Did you try the -Df flag?

zpool import -Df fileserver

Here are some hints: http://docs.huihoo.com/opensolaris/sola ... 04s06.html

Also check if the SATA cables are firmly attached to your drives and controolers and possibly try new ones.
Also post the SMART values here if possible.

[substr]

you might also try again with the corrupted data drive removed from the system.