Hi folks,

 

I have recently added a new disk (disk6) to my Unraid v6.12.5 array. The disk was precleared before adding, and formatted to XFS after adding it; no data was written to it since.

 

Now I have another disk (disk5, XFS) failing with `scsi error badly formed scsi parameters` error, and that disk is full with data. Is there a relatively painless way to take out disk5 and have the newer disk assigned in its slot to rebuild the data from parity? I realize I could have done that with a precleared disk, if only I waited another week before adding it to the array... Now unsure how to proceed, halp.

 

there is no easy way to do quite what you ask.
 

Once you formatted the drive it has data (the empty directory structures) which means it can no longer simply be removed without invalidating parity.  Since invalidating parity would mean you can then not recover the failed disk5 this is not what you want.  The fact there are no files is not relevant.

 

has disk5 been disabled (marked with a red ‘x’)?

 

You are likely to get better informed feedback if you attach your system’s diagnostics zip file to your next post in this thread. It is always a good idea when asking questions to supply your diagnostics so we can see details of your system, how you have things configured, and the current syslog.

 

 

itimpi,

 

Thanks for replying. Yes disk5 is disabled with contents emulated, which suggests that it might or might not be recoverable from parity. Unfortunately I forgot to enable scheduled parity checks and although parity shows as clean, the last time it ran was almost a year ago. Lessons learned here...

 

I realize there might not be an easy-peasy way out of this, I guess what I was asking is if there's a relatively non-painful way to do that. I'm proficient with low level Unixy stuff (pre-Linux even) if that's a consideration. Assuming the parity is correct, can the new disk (disk6) be cleared to zeros while in array? That way I guess I could take it out without ill effects and then swap it for the failed disk and rebuild data on it from parity.

 

I should also add that the data on disabled disk is not über valuable and I can tolerate partial or total loss. What I'm concerned with at this point is not losing parity entirely in case another disk fails soon. That failed 14TB Seagate is one of the batch of 6 that was acquired and installed less than a year ago, not exactly reassuring.

 

I've also attached the diagnostics in case that helps.

 

plex-diagnostics-20241120-1503.zip

Tried to back up some of the data from the failed disk, the few files I copied seem to be not corrupted so there's a chance that parity is actually correct.

 

If it were me, I'd not do any reconfiguration until the data was backed up.

On 11/20/2024 at 7:29 PM, nohuhu said:

Tried to back up some of the data from the failed disk, the few files I copied seem to be not corrupted

Since that is the case, easiest way forward may be to copy from the emulated failed disk directly to the new empty disk.

 

rsync -arv /mnt/disk5/ /mnt/disk6
rsync -narcv /mnt/disk5/ /mnt/disk6

First command does the copy, second command verifies the copy was bit perfect, it only outputs files with differences.

 

After that is complete, you could do a new config and re-order the drives where you want them and rebuild parity.

On 11/20/2024 at 6:14 PM, nohuhu said:

What I'm concerned with at this point is not losing parity entirely in case another disk fails soon.

Short of obtaining another new disk to rebuild the failed one, my method gets you protected the fastest while recovering the data.

 

 

JonathanM,

 

Thanks! This seems like a good idea, I'll try that.