Cache Drive Failing? Trying to remove from Cache pool.

[dirrtyjoe]

It appears my main cache drive is failing. This is a HDD where the other two are SSD so I had planned on removing it anyway. The instructions for removing a drive from a cache pool are numerous and each a bit different. But my goal was to do something like this:

 

[*]Balance Cache

[*]Backup Cache Drive

[*]Stop the array using the webUI

[*]Remove drive to the cache pool

[*]Disconnect power/sata from removed drive

[*]Start the array

[*]Wait for the updated pool to rebalance which will also delete from the pool

 

The initial balance process seems to be hung(?) at: 21 out of about 57 chunks balanced (22 considered),  63% left

 

 

Any insight into how to handle this? Diagnostics attached.

 

[edit] file too big, apparently... here is a download link: https://1drv.ms/u/s!AlrZB2zl8cavgcso3gs6TsVTgw7dvQ

[dirrtyjoe]

Adding some detail - the balance seems to be never ending. I see:

 

Jul 27 11:30:58 Juggernaut kernel: BTRFS info (device sdc1): found 1 extents

Jul 27 11:30:59 Juggernaut kernel: BTRFS info (device sdc1): found 1 extents

Jul 27 11:30:59 Juggernaut kernel: BTRFS info (device sdc1): found 1 extents

Jul 27 11:31:00 Juggernaut kernel: BTRFS info (device sdc1): found 1 extents

Jul 27 11:31:00 Juggernaut kernel: BTRFS info (device sdc1): found 1 extents

Jul 27 11:31:00 Juggernaut kernel: BTRFS info (device sdc1): found 1 extents

 

 

goes on and on - several times per second.

 

Screenshot with drive labels attached.

 

 

[JorgeB]

If one of the cache disks is failing you may not be able to complete the balance, though SMART for the wd looks OK, I can't really tell if all those BTRFS errors are from a bad disk or corrupted filesystem.

[dirrtyjoe]

... I can't really tell if all those BTRFS errors are from a bad disk or corrupted filesystem.

 

As for how to approach correcting it, any suggestions? Wipe the cache drives, reformat, and start over?

[JorgeB]

If it's already back up probably best to create a new pool with the SSDs only, disconnect the other disk and format the SSDs to a different fs (like reiser) before adding them to the new pool, then format to BTRFS again.