ZFS question. Importing degraded RAIDZ2

[somename]

Thanks in advance for any advice I can get. Long story short, I had 6x1TB drives in RAIDZ2 in a system that had internal power surge, and 2 of the RAIDZ2 drive is lost along with the OS(Solaris 11) drive. In a newly installed Solaris 11.1 system, it sees the zpool from failed system with 2 missing drives.
I'm kinda afraid to import the degraded zpool since if there's any corruption on the 4 survived drives, I'd be losing the data. Is there a way to test the integrity of the data without importing the zpool? Would I be able to add the member disks even after I import the pool if I were able to get cloned drives from data recovery company? Thanks for any suggestions/insights.

 

[_Gea]

If you want to test integrity am not aware on another method than import and scrub.Usually you import the pool and do a replace faulted disk -> new disk.

Another option is to import the pool readonly and do a filebased backup prior any repairs (important data first).
If you have a chance to get back a cloned and working disk you may wait

You should think of backups!
ZFS is one of the best available data storage technologies but on a real disaster you need a (external) backup.

 

[HobartTas]

Greetings

Is there a way to test the integrity of the data without importing the zpool?

Not as far as I'm aware, I could be wrong but I think you usually need ZFS running on the opened pool to check the integrity with a scrub, the only other thing you can do is just read all the blocks from start to finish on the remaining drives to check they actually work OK but I wouldn't bother as either the drives work or they don't so this won't tell you anything useful.

This is what I would do.

Firstly, I would say do nothing at this point with the 4 remaining drives.

Secondly, go out and buy some new drives with 4 of them at a minimum but 6 preferably, if this is all that you have to do to recover your data then it will be a lot cheaper than a data recovery service, even if you do need to use the recovery service later on you would probably want to replace your existing drives anyway and just use your existing drives afterwards either for a backup or less sensitive data that you can afford to lose as they may be less reliable after that event. 1TB drives may not be good value for money so perhaps going with 2TB (or even 3TB) might be better.

Image each existing drive to a new one, do not stuff up this procedure for any of your 4 drives or your data is toast

Now try importing the 4 new drives as your pool and do a scrub. Due to the fact that you are effectively running in a Raid-0 configuration any errors on the remaining drives will go below minimum redundancy and you will lose data in that particular stripe (e.g. 2 dead drives + 1 URE) and this means that you will lose the file where the stripe happens to be, fortunately ZFS will tell you which file it is and if you only say have a couple of these types of errors AND you can live with those particular files being lost then you will have succesfully recovered the rest of your pool. Now add the 2 other new drives and let the system re-silver back up to the Raid-Z2 state one drive at a time and you should be OK from this point onwards.

If on the other hand there are files that you can't afford to lose OR on the other hand one of the remaining drives has vast chunks of data lost you will most likely lose a lot of files and in that case you will have to see if you can get one (or both) of the other faulty drives repaired as most likely just the circuit board has fried, if they can be fixed then image them as well when you get them back and then try importing the pool, as long as you have less than 3 damaged blocks per stripe (part of a dead drive OR damaged block(URE)) you will be OK.

That's what I would do, other people may have better opinions and suggest a better method. Also you may want to setup a backup system for your data, another Raid-Z2 pool is my recommendation. I hope this all works out for you.

Cheers