Recover from Cache Pool Drive Failure

[TapIt]

One of the drives in my cache pool suddenly died, thankful it was a raid 1 and I was able to recover the data to a folder on my array following these instructions. Now that I have the data on the array in a temp folder how can I get the cache recovered or get the data into the right spots in the array?

Should I just be copying the data to the array backed shared location for each area? Such as /cache/sharefolder/files -> /array/temp/files -> /array/sharefolder/files? or once I replace the broken drive should I just take the files from the temp directory and put them back on the cache.

leaf-diagnostics-20221103-1905.zip

[JorgeB]

Basically you just need to fix the pool and restore the data to their original folders/shares.

[TapIt]

Thanks for the quick response. Just to confirm the next steps, I should just physically replace the dead NVME drive with a new one. Assign that new one to the slot of the dead one and start the array back up. BTRFS will then repair the raid 1 and the cache should be good to go? 

[JorgeB]

Are you sure the NVMe is dead? It might just have dropped, if you want us to take a look post the diagnostics.

[TapIt]

I’m pretty sure it is dead. I attached the diagnostics to the original post.

 

thanks again.

[JorgeB]

Missed those, but they are before array start, after would be better, SMART does look weird for that device, it's incomplete.

[TapIt]

Attached the diagnostics with the array started. leaf-diagnostics-20221104-1222.zip

[JorgeB]

Other NVMe device is not decrypting, possibly because the LUKS headers are missing, since the pool is backed up you can try re-formatting, then if there are more issues with that device replace it, or if you have a spare replace it now then test the other one.