Failed to import zpool

[dgraybill]

I recently replaced the drive my FreeNAS runs on, with the intention of just importing the ZPool. The pool shows up in the GUI under auto import, so I tried that to no avail. I read a lot online about importing it first in CLI so I tried that numerous times, even using zpool import -f Storage, and I keep getting an error, "one or more devices is currently unavailable. I have not had any problems with the pool in the past, and when I run zpool import, it shows the pool, and all the drives in the pool as "ONLINE". I've been on google for hours now and i'm not finding anyone with the same problem.

Thanks for the help

 

[Stephens]

I can't answer your question, but please provide your FreeNAS version for those who can.

There's a process documented in the documentation for replacing drives. Did you follow it? Without even reading it myself, seems to me you're supposed to offline the drive, replace it, and on-line it. You can read for yourself what the exact process is.

 

[paleoN]

I recently replaced the drive my FreeNAS runs on, with the intention of just importing the ZPool.

Do you have the original and does it still work?

I read a lot online about importing it first in CLI so I tried that numerous times, even using zpool import -f Storage, and I keep getting an error, "one or more devices is currently unavailable.

It helps to actually understand the commands you are running. ;)

I would use 8.3.0-RC1 to attempt recovery as it has the improved ZFS v28 code in it. Burn that to a new USB drive and try from there. No -f imports unless you know what you are doing.

 

[dgraybill]

Well I already typed this once, but it didn't seem to post. It was version 8.0.3. I had re used the disk I originally used. I do not still have the original, the drive died. I tried to import using 8.3 but I got the same output. I've attached a photo of the output:
https://lh3.googleusercontent.com/-...AAAAAC8/bzEUyGAjLi4/s1152/20121023_175026.jpg

 

[paleoN]

Too bad. I was hoping you could pull the original zpool.cache file off. You sure you didn't have another disk in there somewhere: a stripe, log or cache? For each disk/partition in the pool:

Code:

zdb -l /dev/adaXp2

Assuming they all match each other post one of them.

 

[dgraybill]

They do not all seem to match. Here is the output:



--------------------------------------------
LABEL 3
--------------------------------------------
version: 15
name: 'Storage'
state: 0
txg: 6833
pool_guid: 17646329889008502608
hostid: 1956700126
hostname: ''
top_guid: 8476132071472504239
guid: 8476132071472504239
vdev_tree:
type: 'disk'
id: 0
guid: 8476132071472504239
path: '/dev/ada0p2'
whole_disk: 0
metaslab_array: 26
metaslab_shift: 33
ashift: 9
asize: 998052462592
is_log: 0
[root@freenas ~]# zdb -l /dev/ada0p2





--------------------------------------------
LABEL 3
--------------------------------------------
version: 15
name: 'Storage'
state: 0
txg: 6833
pool_guid: 17646329889008502608
hostid: 1956700126
hostname: ''
top_guid: 5895470883330402436
guid: 5895470883330402436
vdev_tree:
type: 'disk'
id: 1
guid: 5895470883330402436
path: '/dev/ada1p2'
whole_disk: 0
metaslab_array: 23
metaslab_shift: 34
ashift: 9
asize: 2998440558592
is_log: 0
[root@freenas ~]# zdb -l /dev/ada1p2






--------------------------------------------
LABEL 3
--------------------------------------------
version: 15
name: 'Storage'
state: 0
txg: 6833
pool_guid: 17646329889008502608
hostid: 1956700126
hostname: ''
top_guid: 2720693610616754064
guid: 2720693610616754064
vdev_tree:
type: 'disk'
id: 2
guid: 2720693610616754064
path: '/dev/ada2p2'
whole_disk: 0
metaslab_array: 69
metaslab_shift: 31
ashift: 9
asize: 397936164864
is_log: 0
[root@freenas ~]# zdb -l /dev/ada2p2

 

[dgraybill]

--------------------------------------------
LABEL 3
--------------------------------------------
version: 15
name: 'Storage'
state: 0
txg: 6833
pool_guid: 17646329889008502608
hostid: 1956700126
hostname: ''
top_guid: 8476132071472504239
guid: 8476132071472504239
vdev_tree:
type: 'disk'
id: 0
guid: 8476132071472504239
path: '/dev/ada0p2'
whole_disk: 0
metaslab_array: 26
metaslab_shift: 33
ashift: 9
asize: 998052462592
is_log: 0
[root@freenas ~]# zdb -l /dev/ada0p2

 

[dgraybill]

--------------------------------------------
LABEL 3
--------------------------------------------
version: 15
name: 'Storage'
state: 0
txg: 6833
pool_guid: 17646329889008502608
hostid: 1956700126
hostname: ''
top_guid: 5895470883330402436
guid: 5895470883330402436
vdev_tree:
type: 'disk'
id: 1
guid: 5895470883330402436
path: '/dev/ada1p2'
whole_disk: 0
metaslab_array: 23
metaslab_shift: 34
ashift: 9
asize: 2998440558592
is_log: 0
[root@freenas ~]# zdb -l /dev/ada1p2

 

[dgraybill]

--------------------------------------------
LABEL 3
--------------------------------------------
version: 15
name: 'Storage'
state: 0
txg: 6833
pool_guid: 17646329889008502608
hostid: 1956700126
hostname: ''
top_guid: 2720693610616754064
guid: 2720693610616754064
vdev_tree:
type: 'disk'
id: 2
guid: 2720693610616754064
path: '/dev/ada2p2'
whole_disk: 0
metaslab_array: 69
metaslab_shift: 31
ashift: 9
asize: 397936164864
is_log: 0
[root@freenas ~]# zdb -l /dev/ada2p2

 

[paleoN]

That's a dumb pool setup of, at least, a 3 disk stripe. I was thinking it was a 3 x disk raidz1. You never answered my other question.

You sure you didn't have another disk in there somewhere: a stripe, log or cache?

The labels appear fine to me aside from not being in [code][/code] tags, which makes them harder to read.

 

[dgraybill]

I do not. The only other disk was the old os disk and it completely died.

 

[William Grzybowski]

Try with -F:

# zpool import -fF -R /mnt Storage

 

[dgraybill]

Shell
[root@freenas ~]# zpool import -f -r /mnt/Storage
cannot import '/mnt/Storage': no such pool available
[root@freenas ~]# zpool import -f -r Storage
cannot import 'Storage': one or more devices is currently unavailable
[root@freenas ~]#

 

[William Grzybowski]

No, that is not that I told you to type, it was easy as ctrl+c+v:

# zpool import -fF -R /mnt Storage

It is case-sensitive, there is a space between /mnt and Storage

 

[dgraybill]

Sorry, I was using RDP on my cell phone driving across Nebraska, I misread what you typed, and I couldn't copy paste because my android doesn't share its clipboard with my RDP program. It ran that command for approximately 2 hours, and when I came home my shell window was closed so I can't give you the exact output, but it was something to the effect of ignoring 37 seconds of transmission, and that 'storage' returned to its previous state. Now, if I go to auto import in the GUI, storage is no longer listed there, and when I run 'zpool import' from the CLI, storage is no longer listed there either.

 

[William Grzybowski]

Because it imported...

# zpool export Storage

now go to the gui and use Auto Import

 

[dgraybill]

So far so good, now as long as I don't lose any more OS drives. Guess I'll set it up to back up somewhere. What exactly does ignoring 37 seconds of transmission mean?

 

[dgraybill]

WARNING: The volume Storage (ZFS) status is UNKNOWN: One or more devices has experienced an error resulting in data corruption. Applications may be affected.Restore the file in question if possible. Otherwise restore the entire pool from backup.

This shows in the GUI, but all the drives show as online.

 

[cyberjock]

If everything is fine with your ZFS pool, a reboot will likely fix the issue. I have previously posted that there's some inconsistency on when the ZFS status actually changes. I'd recommend you do a reboot.

 

[dgraybill]

I tried that, and it still shows the same. I can read and write data so I don't know what it's so upset about.