Remove drive from Cache pool

[kricker]

 

A while back I wanted to replace my cache drive. Instead of replace the one I had, I added a second and now have a Raid 1 cache pool.

Data, RAID1: total=15.00GiB, used=13.52GiB
System, RAID1: total=32.00MiB, used=16.00KiB
Metadata, RAID1: total=1.00GiB, used=9.58MiB
GlobalReserve, single: total=16.00MiB, used=112.00KiB

I want to now drop back down to one single cache drive as the smaller drive is having pre-fail issues. I am trying to follow the following directions to drop back down.

But I am stuck at the Balance operation. it has been running for a couple hours and is still reporting the following:

Balance on '/mnt/cache' is running
3 out of about 4 chunks balanced (17 considered),  25% left

Is it normal for that to take so long?

[JorgeB]

6 minutes ago, kricker said:

Is it normal for that to take so long?

 

Not for less than 15GiB of data, post your diagnostics.

[kricker]

Here they are.

tower-diagnostics-20171113-1433.zip

[JorgeB]

Looks like it's stuck, also it's continuing after the previous reboot:

 

Quote

Nov 13 11:52:23 Tower emhttp: shcmd (58): mkdir -p /mnt/cache
Nov 13 11:52:23 Tower emhttp: shcmd (59): set -o pipefail ; mount -t btrfs -o noatime,nodiratime -U cefcc342-daaf-4c80-9fa0-34d00f89b66c /mnt/cache |& logger
Nov 13 11:52:23 Tower kernel: BTRFS info (device sdd1): disk space caching is enabled
Nov 13 11:52:23 Tower kernel: BTRFS info (device sdd1): has skinny extents
Nov 13 11:52:24 Tower kernel: BTRFS info (device sdd1): checking UUID tree
Nov 13 11:52:24 Tower kernel: BTRFS info (device sdd1): continuing balance

 

Try rebooting one more time, if there's still important data on the pool back it up before rebooting, if still stuck after reboot, remove the extra device and reformat.

[kricker]

All that is on the pool is the docker image and the appdata, which I already backed up.

[kricker]

Still stuck after a reboot. Guess I will just have to pull it and hope for the best

 

[kricker]

So now I am stuck in a loop. Every time I reboot the balance process starts again and never finishes. So I forced the server down and pulled the drive I no longer want. Now it boots up and shows that drive as missing but is still running the balance operation. Clicking on cancel will not stop it either. It just hangs the webGUI. At next reboot it all starts again. Trying to stop the array causes the system to also hang and I have to force it down.

[JorgeB]

Disable array auto-start, reboot, and re-format the disk by changing the filesystem.

[kricker]

I forced it down, re-booted in safe mode and adjusted the cache drive settings and re-formatted. I have to re-do all my dockers, but I am back up and working properly now.

Edited November 14, 2017 by kricker