[OpenIndiana-discuss] Help on Rescue Pool

[Nikola M]

On 06/ 3/15 06:54 AM, Handojo via openindiana-discuss wrote:
> The system boots into rpool correctly, but /datapool is unavailable.
>
> ------------------------
> pool: datapool
>   state: UNAVAIL
> status: One or more devices could not be opened. There are insufficient replicas for the pool to continue functioning
> action: Attach the missing devices and online it using 'zpool online'
>
> config:
>
> 	NAME        STATE     READ WRITE CKSUM
> 	datapool  UNAVAIL          0           0            0  Insufficient replicas
> 	  raidz1    UNAVAIL          0           0            0  Insufficient replicas
> 	    c9d0    OFFLINE           0           0            0
> 	    c7d1    UNAVAIL          0           0            0  cannot open
>
> ------------------------

Raidz could be looked at as RAID5, it requires 3 devices to function 
properly.
It can withstand loosing one device durign operation but you need to 
give it 3 devices as I understand.
Here it seems it did not recognize disks at all.

Are you still on Opensolaris 2009.06(111b) with rpool? You can migrate 
install to latest snv_134 from Openindiana /legacy repo and after that 
you can update to Openindiana /dev (a3, then a7 and up to a8/a9)
Moving to testing Hipster requres separate install ATM (I did id by 
installing in VBox and zfs send/receive to machine's new emptied Boot 
Environment)

> After this, I issue :
>
> --------------------------------------
> %zpool destroy datapool
>
> %zpool create -f -m /datapool datapool /dev/rdsk/c14t2d0
>
> --------------------------------------
>
> and the datapool is created, and all the directory and files are gone.
>
> How do I undo this, or how do I fix this ?
>

Well you killed it, Jim. It's gone now.
After destroying it, it is gone as I understand.

I just hope it's a new disk , if it is an old one that was member of the 
pool called the same name,
then maybe you could visit your local data restoration shop.
You can maybe try doing something with the zdb of hopingly if new pool 
is created on new disk and not in the old one, and you have ate least 2 
devices left from raidz pool, you can still bring it back.
Or bring data from the backup (RAIDs and replications are no excuse for 
not doing backups)

If you want to ensure that in some time in the future you try again with 
restoring data from destroyed pool,
I would do 'dd' of whole devices to a vdevs/image files, compressed.
That way you can share it maybe with mega.co.nz or similar sharing site 
and allow someone else try to do a recovery, not waiting for your local 
restoration shop.

Also, try asking on illumos-zfs mailing list, too,
http://wiki.illumos.org/display/illumos/illumos+Mailing+Lists
(And on other ZFS related lists or freenode irc channels, OpenZFS is 
available also on ZFSOnLinux and OSX etc)