BTRFS errors on a mirrored cache pool, now shows no filesystem. OMG HELP

[dRuEFFECT]

I have 2 500gb crucial ssd's mirrored in a raid 1 cache pool. I keep everything updated, on the latest 6.9.2.

 

Everything was swimming along nicely until this morning when I updated Fix Common Problems, then went to CA and showed me that docker was disabled. I check things out and see that my cache drive is in read only mode, and one of the 2 SSDs was showing BTRFS errors to trigger this. I was confused why I would have any issues, if one drive had a problem, wouldn't the second be able to keep things afloat?

 

So I tried stopping the array, this way i can add a spare SSD to pool devices and copy the cache drive contents, then possibly drop BTRFS for XFS and run a single cache drive with scheduled backups. Stopping the array was getting hung up on unmounting the cache drive. i tried manually unmounting, but no joy. So I rebooted into safe mode and now both SSDs are showing as unmountable with no file system.

 

For the love of god, please tell me there's a way to recover the cache drive contents. I've been running a mirrored cache pool solely for the peace of mind that i have redundancy, how could this be happening? What can i do?

 

[JorgeB]

Please post the diagnostics.

[dRuEFFECT]

4 minutes ago, JorgeB said:

Please post the diagnostics.

Attachedandrews-unraid-diagnostics-20220228-1258.zip

[JorgeB]

Mirror is for redundancy, it won't help with filesystem corruption which is the problem here, there are some recovery options here, see if you can backup anything important then re-format the pool and restore the data.

[dRuEFFECT]

50 minutes ago, JorgeB said:

Mirror is for redundancy, it won't help with filesystem corruption which is the problem here, there are some recovery options here, see if you can backup anything important then re-format the pool and restore the data.

THANK YOU!!!

 

btrfs restore is doing the trick for me

Edited February 28, 2022 by [email protected]

[dRuEFFECT]

51 minutes ago, JorgeB said:

Mirror is for redundancy, it won't help with filesystem corruption which is the problem here

 

for my new cache pool, would it make sense to run a single drive primary cache and schedule rsync on the entire drive to a secondary SSD pool, this way both drives are XFS, both drives have independent filesystems, and i could stop the array and hot swap the backup cache in case of failure. is this plausible or am i missing something?

[JorgeB]

For backups it would be much better, assuming the sync is done frequently enough.

[dRuEFFECT]

roger that. thanks for your help. i almost had an anxiety attack

[dRuEFFECT]

edit: i re-ran it and yea it appears to skip existing files pretty quickly

 

 

btrfs restore ran for a while but hit a not enough memory error.. not sure what to do next

 

would re-running skip over what's already restored?

 

Edited February 28, 2022 by [email protected]

[JorgeB]

12 hours ago, [email protected] said:

would re-running skip over what's already restored?

Not sure, you can try, or restore to a different path.