Both of the drives in my mirror has critical issues

one drive works great but makes a horrible grinding like sound example and the other has went offline i dont know when but probably not very long ago .
when i do “zpool import poolname” it says the drive has an I/O error but i can still mount it but will result in data loss.
while the other drive has a copy of the pool in perfect condition i dont know if its worth to risk the data and drive (until i can replace it) right now when the second drive has a bad copy.
i do have enough drives that work (at least i think) to backup my pool to but its 1tb drives 1 is a different make and model and the others are 500gb that are also different make and model and all of the drives are pushing on 7 years now.
what do i do in this situation?
both of the drives in my pool are WD40EZAX still under warranty as they are less then a year old

full zpool import error

cannot import 'storage' : I/O error
       Recovery is possible, but will result in some data loss.
        Returning the pool to its state as of Wed Jul  3 12:17:17 2024
        should correct the problem.  Approximately 19 seconds of data
        must be discarded, irreversibly.  After rewind, at least
        one persistent user-data error will remain.  Recovery can be attempted
        by executing 'zpool import -F storage'.  A scrub of the pool
        is strongly recommended after recovery.

You unfortunately did not provide much data.

First, the sound is of the drive performing a lot of seeking action. That is fine IF the drive is running a test that performs a lot of seeks, however I suspect you are not doing that. It definitely sounds like a drive failure.

Please provide the output of the following commands:
zpool status -v storage and it may not result in anything much based on the message above.
smartctl -x /dev/sda for each drive. /dev/sdb and /dev/sdc. One of these should be your boot drive, the others your mirror drives. This data will provide us a lot of information. It will also provide the period of time a Long test should take for each drive.

After posting that data run smartctl -t long /dev/sda and for all the drives you have. The tests will run until they complete or fail.

Using the data from the first set of smartctl commands wait the maximum amount of time for the drives to complete the testing and then run the first set of commands smartctl -x /dev/sda and for the other, post the output of each.

That is not what the warning message says.

I would not recommending running the zpool import -F storage command yet, wait until we can evaluate the results. You may end up doing that anyway but none of us want someone to lose data.

Run the tests and certain errors will validate the RMA.

If you have a place you can copy your data to, you should do that, but not by replacing a drive in your machine. You could add a drive as a new pool and copy there if that is the only option you have, but do not replace the original drives yet.

Best of luck to you.

2 Likes

What they said, your first order of business is copying off any important data first to a safe place before doing anything.