Zpool on GELI halts during import

Chreo chreo at chreo.net
Wed Dec 23 17:16:01 UTC 2009


Hi there,

I've a raidz zpool on 6 1,5TB vdevs running on GELI GEOMs. It has been 
performing fine until a few days ago when all of a sudden it started to 
fail being imported. This is on 8-stable

zpool import reports:
zpool import                                       
  pool: Ocean
    id: 18338821095971722517
 state: ONLINE
status: One or more devices contains corrupted data.
action: The pool can be imported using its name or numeric identifier.
   see: http://www.sun.com/msg/ZFS-8000-4J
config:

    Ocean                 ONLINE
      raidz1              ONLINE
        label/Disk_1.eli  ONLINE
        label/Disk_2.eli  ONLINE
        label/Disk_3.eli  ONLINE
        label/Disk_4.eli  UNAVAIL  corrupted data
        label/Disk_5.eli  UNAVAIL  corrupted data
        label/Disk_6.eli  ONLINE

but doing the actual import simply cause the process to halt it's 
progress (at least a ktrace does not reveal any activity).

I've also tried doing the import using:
zpool import -f -o failmode=continue Ocean
with the same result.

Right, so my first thought was that this has been caused by the 
notorious uberblock issue (where the . Unfortunately doing
zdb -uuv -e Ocean
also stops responding (no disk activity noted after 1s of running the 
command).

zdb -l /dev/label/Disk_X.eli
Reports identical data for all 6 drives (they only differ in the GUID as 
expected).

Now running on 7-stable instead (the system was upgraded a monthe ago) 
does not change the behaviour fundamentally, import still stops 
responding, however zdb -e Ocean now cause a segfault so it's not much 
of an "improvment".

Any ideas as to how I could get this into a importable state or find out 
what went backwards?

Cheers
Christian Elmerot





More information about the freebsd-fs mailing list