Cache Drive Stopped Working

[kenoka]

I had a cache pool of three drives. A couple of days ago, one of the drives dropped out of the pool, and I had to rebuild my dockers. I tested the suspect drive, and it seemed fine, but I had been using it in an external eSATA dock before putting it into my case, and it seems like that dock has failed. So I put the drive into the case, and started up.

 

The drive showed up again, and shows as part of the cache pool, but the cache capacity remains at 480 GB, rather than ~720 GB (two 480 GB drives and one 500 GB). The third cache drive will also show as spun down after some time, and there is essentially no read or write activity on it. When I originally added it, it was quite busy, and I expected this to happen again.

 

SMART reports attached.

mediaserver-smart-20161017-0737.zip

mediaserver-smart-20161017-0737_long.zip

[JorgeB]

You'll need to clear the SSD that dropped out before re-adding it, all data on it will be lost.

 

-if there's important data on the cache pool backup in case something goes wrong.

-stop array

-unassign the problem SSD

-clear it by typing on the console:

 

blkdiscard /dev/sdX

replace X with actual letter.

 

-start the array with the only two remaining cache devices, make sure cache pool mounts and on the cache page "btrfs filesystem show" total devices are correct and it's not displaying "***some devices missing"

-stop array, re-add SSD to the pool, a balance should begin, when done check again "btrfs filesystem show" and that total devices are correct.

[kenoka]

When I restarted the array, the two drive cache pool shows as unmountable. I'm sorry, I don't know where the cache page is.

 

Edit: Never mind, I just had to stop and start again. Thanks very much! It seems to be re-balancing now.