ZFS Pool errors when importing on 9.1.0

[trevorault]

I've been running FreeNAS for several years and recently had issues accessing the GUI and AFP/NFS shares. I installed 9.1.0 on a brand new USB stick and have errors with my ZFS pool. Do I need to replace drives first (it looks like the 2nd has failed with the 1st one having issues as well) or do something with the existing pool to get it online first? I looked through the files with errors and am OK if I lose them. What are the recommended next steps? I appreciate any help.

Code:

[root@freenas] ~# zpool status
  pool: Carlos
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://illumos.org/msg/ZFS-8000-8A
  scan: resilvered 6.95G in 0h9m with 25724 errors on Fri Aug 23 07:00:37 2013
config:
 
    NAME                                            STATE    READ WRITE CKSUM
    Carlos                                          DEGRADED    0    0 34.3K
      raidz1-0                                      DEGRADED    0    0 95.4K
        gptid/b2600453-714d-11e1-90bd-00151731e8a6  ONLINE      0    0  186
        gptid/b2afe3dd-714d-11e1-90bd-00151731e8a6  FAULTED    47 1.32K    28  too many errors
        gptid/b2feb193-714d-11e1-90bd-00151731e8a6  ONLINE      0    0    0
        gptid/b34e3de7-714d-11e1-90bd-00151731e8a6  ONLINE      0    0    0
 
errors: 25725 data errors, use '-v' for a list

 

[cyberjock]

First, read the link in my sig on why RAIDZ1 is just stupid(as you have become yet another example of why RAIDZ1 is really no redundancy at all).

Then, after reading that link you'll understand why UREs are bad and why you best bet is to create a new pool with RAIDZ2.

Unless you delete all of the offending files, you will always have a list of data errors. And if any of those errors are metadata, you practically have to recreate the pool from scratch to get rid of them unless you know exactly what file is attached to that metadata(nobody I know has figured out how to do that).

Personally, if you are like everyone else in the forum, you are looking at rebuilding your pool from scratch after you copy off whatever files you can salvage if you ever want the pool to be "healthy" again.

 

[trevorault]

Ok, I will read up on RAIDZ2 and consider rebuilding from scratch. What do I need to do next to recover the files? I only have a couple months of files that are not backed up. Do I need to replace the failed drive first or is there a way to recover files before then?

 

[cyberjock]

Totally up to you, but remember that you have to destroy the zpool to recreate the vdev with the correct type. So you are looking at copying everything off the pool, destroying and recreating the pool, then copying all of the data back.

At this point the damage is done. The real question is how long until you end up with an unmountable pool. Judging from your situation, probably less time than you want to know. I'd just start copying your data off now before you lose a second drive and lose everything. There are no "recovery tools" for zpools, so if the pool becomes unmountable you lose all data on the pool permanently.

 

[Yatti420]

I got lucky.. Had a lot of problems awhile back.. The data literally started eating itself as I accessed my bad drive.. causing mass checksum errors etc.. No ECC.. So I lost some files.. The drive was bad and needed to be RMA (the WD20EARS)..