Options For Double Drive Failure Raidz1?

[Fuzzball]

Background:
So I know better, but just like many I spend most of my time fixing other people's stuff rather than maintaining my own and I never got around to implementing my backup solution. I have two drives that failed close to each other and I didn't have enough time to replace the first failure. Some data was written after the first drive failure. I have the drives offline for now.

Drive1 seems like its controller failed; it doesn't seem to spin up or move the heads. Drive2 has the newly written data and it sounds like the bearings have gone bad with the actuator arm clicking three times before giving up.

Question:
If I can get Drive1 working with Drive2's board (if that's the actual problem), would it be possible to get it working in the pool? I'm fine with loosing the newly written data as I just need some of the old data. There are no snapshots unless ZFS automatically took one when it removed Drive1 from the pool.

If not, I'll try and cannibalize Drive1 to get Drive2 working.

 

[Sebulon]

@Fuzzball

I say go for it, sounds like a decent plan. Good luck, hope it works! Regardless, get a backup aftwards

/Sebulon

 

[peter@]

I've done this in the past and have had some tremendously good luck. But don't kid yourself, whether this works largely a factor of care + luck.

If the drives have the same firmware on their controllers you're in much better shape than slightly different versions. Some drives store their operating firmware on the drives themselves.. ie: they bootstrap a tiny system to read the real firmware from the drive.

ns0.freebsd.org was salvaged this way once. The drive failed to spin up after the cluster's colo move a few years back and I was lucky enough to find a 10 year old equivalent drive + controller that I could savlage the board from.

There were lots of read errors (presumably the bad sector maps were out of sync) but I got all the important data off the machine.