Pool is in a DEGRADED state (GPT Table corrupt)

[mainstream]

Hello everyone,

I have a question regarding my drives. A couple days ago my PSU short-circuited, after replacement my "HDD" pool was degraded. I did a couple scrubs and cleared the errors.

The drives are back online without errors, also the S.M.A.R.T. test (long) are SUCCESS, but when i run fdisk -l:

How can i fix this gpt table?

Best regards

 

[sretalla]

There are a few ways to address it.

If that disk is in a pool with sufficient redundancy, you can offline the disk, wipe it and replace it with itself.

If it's the only disk in the pool, then you may need to copy over the partition table from another similar disk or recreate the table in-place... I advise looking up the detailed steps as that can go badly wrong if you don't do it right.

 

[mainstream]

It's has redundancy (raidz2). So you mean to disconnect it, hook it up to another computer and completely wipe it and resilver it?

 

[Patrick M. Hausen]

If it's the only disk in the pool, then you may need to copy over the partition table from another similar disk or recreate the table in-place... I advise looking up the detailed steps as that can go badly wrong if you don't do it right.

FreeBSD's gpart has got an easy to use recover command to fix corrupt GPT tables. Isn't there an equivalent on Debian?

 

[mainstream]

I guess gdisk?

Strange that it mentions a valid GPT, while fdisk outputs an error.

Edit: Ah, i see that the main GPT table is valid, but the backup GPT table is corrupt. Recovery (R) would try to recover from the backup.