[SOLVED] How to convert Cache pool to single mode

[bcabalka]

My brain doesn't appear to be working right now and when I go to the cache pool and convert to single mode, it doesn't appear to expand the space available on the pool. Is there a way to get this converted while retaining the data on the pool?

 

I have 2x 1TB SSDs and was hoping to use all available space.

 

 

Edited December 31, 2022 by bcabalka

[JorgeB]

Select "convert to single profile" and click balance.

[bcabalka]

I have done that, but it does not do anything. Any other ideas? Can I run something from terminal?

[bcabalka]

Looking at the syslog, it appears there's a BTRFS error. That would explain why it isnt converting.

 

Dec 30 08:09:37 cortex  ool www[13111]: /usr/local/emhttp/plugins/dynamix/scripts/btrfs_balance 'start' '/mnt/cache' '-dconvert=single,soft -mconvert=raid1,soft'
Dec 30 08:09:37 cortex kernel: BTRFS info (device sdb1): balance: start -dconvert=single,soft -mconvert=raid1,soft -sconvert=raid1,soft
Dec 30 08:09:37 cortex kernel: BTRFS info (device sdb1): relocating block group 2442826612736 flags data|raid1
Dec 30 08:09:37 cortex kernel: BTRFS warning (device sdb1): csum failed root -9 ino 299 off 969187328 csum 0x3d995087 expected csum 0x1930d4f7 mirror 1
Dec 30 08:09:37 cortex kernel: BTRFS error (device sdb1): bdev /dev/sdc1 errs: wr 0, rd 0, flush 0, corrupt 887, gen 0
Dec 30 08:09:37 cortex kernel: BTRFS warning (device sdb1): csum failed root -9 ino 299 off 969187328 csum 0x3d995087 expected csum 0x1930d4f7 mirror 2
Dec 30 08:09:37 cortex kernel: BTRFS error (device sdb1): bdev /dev/sdb1 errs: wr 0, rd 0, flush 0, corrupt 726, gen 0
Dec 30 08:09:37 cortex kernel: BTRFS warning (device sdb1): csum failed root -9 ino 299 off 969187328 csum 0x3d995087 expected csum 0x1930d4f7 mirror 1
Dec 30 08:09:37 cortex kernel: BTRFS error (device sdb1): bdev /dev/sdc1 errs: wr 0, rd 0, flush 0, corrupt 888, gen 0
Dec 30 08:09:37 cortex kernel: BTRFS warning (device sdb1): csum failed root -9 ino 299 off 969187328 csum 0x3d995087 expected csum 0x1930d4f7 mirror 2
Dec 30 08:09:37 cortex kernel: BTRFS error (device sdb1): bdev /dev/sdb1 errs: wr 0, rd 0, flush 0, corrupt 727, gen 0
Dec 30 08:09:38 cortex kernel: BTRFS info (device sdb1): balance: ended with status: -5

 

[JorgeB]

Please post the complete diagnostics.

[bcabalka]

56 minutes ago, JorgeB said:

Please post the complete diagnostics.

Attached

cortex-diagnostics-20221230-0936.zip

[JorgeB]

Btrfs is detecting data corruption on both devices:

 

Dec 29 21:24:42 cortex kernel: BTRFS info (device sdb1): bdev /dev/sdb1 errs: wr 0, rd 0, flush 0, corrupt 680, gen 0
Dec 29 21:24:42 cortex kernel: BTRFS info (device sdb1): bdev /dev/sdc1 errs: wr 0, rd 0, flush 0, corrupt 837, gen 0

 

This suggests a RAM or other hardware problem, start by running memtest, once that's resolved you should run a scrub, any uncorrectable files should be deleted/restored from a backup, then reset the stats and keep monitoring.

[bcabalka]

Memtest showed errors, started by reseating modules.

 

Ran scrub, found csum error files, removed, re-ran scrub, came up clean.

 

Reset stats and set up user script to notify of future errors.

 

Successfully converted to single mode.

---

Thanks again for your guidance.