Hello and Happy New Year,
Following a power failure I have been unable to mount my ZFS raidz pool. I was able to export the pool, but subsequent "import -f" attempts have failed. I am looking for advice on how to correct the corruption and get the pool back.
CONFIGURATION: I am running FreeBSD 8.2-RELEASE-p3 as an ESXi virtual machine. I am ready and able to switch to a different FreeBSD or OpenSolaris version for the purpose of this troubleshooting. The raidz pool is comprised of 8 x 2TB drives. Each disk has an empty 2GB slice (p1), followed by a 1.8TB slice (p2) for use by ZFS.
PROBLEM: Following a brief power-off of the hard disks' enclosure, the pool cannot be mounted.
The output of "#gpart list" is too long for this post. I will post that output in an immediate follow-up.
I've researched the problem myself for a while. The TS here had the same problem resulting from a power failure, and the same zpool status. The advice given was that it was likely an error with "labeling, dangling device links, or overlapping partitions." Based on that info, TS reported he fixed the problem:
I am writing here because I am still a novice at FreeBSD. I know only enough about labels to know they should not be messed with lightly. I know nothing about "dangling device links" and "overlapping partitions".
Any advice you can offer about how to proceed in a manner that won't destroy my 14TB pool is greatly, greatly appreciated.
Thank you very much in advance.
Following a power failure I have been unable to mount my ZFS raidz pool. I was able to export the pool, but subsequent "import -f" attempts have failed. I am looking for advice on how to correct the corruption and get the pool back.
CONFIGURATION: I am running FreeBSD 8.2-RELEASE-p3 as an ESXi virtual machine. I am ready and able to switch to a different FreeBSD or OpenSolaris version for the purpose of this troubleshooting. The raidz pool is comprised of 8 x 2TB drives. Each disk has an empty 2GB slice (p1), followed by a 1.8TB slice (p2) for use by ZFS.
PROBLEM: Following a brief power-off of the hard disks' enclosure, the pool cannot be mounted.
Code:
#zpool import
pool: ZFSstore
id: 12422265870499905405
state: FAULTED
status: The pool metadata is corrupted.
action: The pool cannot be imported due to damaged devices or data.
The pool may be active on another system, but can be imported using
the '-f' flag.
see: http://www.sun.com/msg/ZFS-8000-72
config:
ZFSstore FAULTED corrupted data
raidz1 ONLINE
da1p2 ONLINE
da2p2 ONLINE
da3p2 ONLINE
da4p2 ONLINE
da5p2 ONLINE
da6p2 ONLINE
da7p2 ONLINE
da8p2 ONLINE
Code:
#zpool import -f ZFSstore
internal error: Illegal byte sequence
Abort
Code:
#zdb
ZFSstore
version=15
name='ZFSstore'
state=0
txg=1219414
pool_guid=12422265870499905405
hostid=2624045851
hostname=''
vdev_tree
type='root'
id=0
guid=12422265870499905405
children[0]
type='raidz'
id=0
guid=9661925766897747075
nparity=1
metaslab_array=23
metaslab_shift=37
ashift=9
asize=15985973133312
is_log=0
children[0]
type='disk'
id=0
guid=1621383351031684258
path='/dev/da1p2'
whole_disk=0
DTL=102
children[1]
type='disk'
id=1
guid=4244177242659624564
path='/dev/da2p2'
whole_disk=0
DTL=101
children[2]
type='disk'
id=2
guid=14313008757061126085
path='/dev/da3p2'
whole_disk=0
DTL=100
children[3]
type='disk'
id=3
guid=8151646539787060344
path='/dev/da4p2'
whole_disk=0
DTL=99
children[4]
type='disk'
id=4
guid=12023942581224222406
path='/dev/da5p2'
whole_disk=0
DTL=98
children[5]
type='disk'
id=5
guid=4292570899629833601
path='/dev/da6p2'
whole_disk=0
DTL=97
children[6]
type='disk'
id=6
guid=14307962413307638883
path='/dev/da7p2'
whole_disk=0
DTL=96
children[7]
type='disk'
id=7
guid=5832102164715480864
path='/dev/da8p2'
whole_disk=0
DTL=95
The output of "#gpart list" is too long for this post. I will post that output in an immediate follow-up.
I've researched the problem myself for a while. The TS here had the same problem resulting from a power failure, and the same zpool status. The advice given was that it was likely an error with "labeling, dangling device links, or overlapping partitions." Based on that info, TS reported he fixed the problem:
I relabeled the disks using the partition information I was able to get from the FreeBSD Live CD, and then was able to import/repair the zpool using the latest OpenSolaris Live CD.
I am writing here because I am still a novice at FreeBSD. I know only enough about labels to know they should not be messed with lightly. I know nothing about "dangling device links" and "overlapping partitions".
Any advice you can offer about how to proceed in a manner that won't destroy my 14TB pool is greatly, greatly appreciated.
Thank you very much in advance.