Replaced 2 of 4 disks in RZ2 now can't get them back into pool

[dclutter1]

I have a raidz2 with 4 2tb drives. 2 of them went bad a few weeks ago and I replaced them with new drives.

Now I'm not sure what to do with the 2 new drives. They're recognized everywhere in the gui, zfs info, etc., but I'm not sure how to add them back. As I understand it, I can't add disks to a current virtual device (right?). If I create a new virtual device with the 2 new disks, I won't have a raidz2 right?

Do I need to just backup and rebuild the vdev/pool?



I made a thread about this in the disk/controller section but it has kind of died off, plus I feel like it may fit into this section better now anyway.

[raulfg3]

FAQ 149: http://wiki.nas4free.org/doku.php?id=faq:0149

[dclutter1]

Right. I did that.

Now I have 2 new disks NOT associated with a vdev or a pool. What do I have to do to get them back into the raid?

[raulfg3]

Go to Disks|ZFS|Configuration|Synchronize and post screen capture of your detectec and real config

[dclutter1]

Thank you

[raulfg3]

sorry , not sure what can be next step, wait until other power user can help you.

[b0ssman]

i am still waiting for you to post the smart values of your drive.

[dclutter1]

i am still waiting for you to post the smart values of your drive.

Why?

Don't you think it's considerably more important to re-establish redundancy before worrying about that?

[b0ssman]

because it looks like your other drives are dying as well and might have damaged the pool. then there is no point is trying to save the pool.

[dclutter1]

So since there's no way to add the 2 new disks back in, best bet is to rebuild the whole thing?

I began backing everything up from the beginning because I thought this might be the case.

[m_seitz]

I assume you already solved your problem. But would it have been possible to clone the old failed drives onto the new ones? Usually, a failed drive has bad sectors or will only run with cooling (inside a freezer!), but most of the data can be accessed.
Cloning failed drives and putting the clones back into a raid should make things easier and minimise data loss, in case another drive fails during resilvering. The chances that the same blocks become inaccessible on several HDDs should be slim. Therefore cloning (if possible) could theoretically retain all data, even if every single HDD in a raid fails.
Or am I missing something here?