A lot of BTRFS warnings

[Shin Zearim]

 I am noticing a lot of Btrfs errors going on from one of my cache drives. Is this a failure of the drive? I attached my diagnostics. 

vengeance-diagnostics-20181114-1123.zip

[JorgeB]

Nov 11 22:27:00 Vengeance kernel: BTRFS info (device sdg1): bdev /dev/sdg1 errs: wr 24652993, rd 16582576, flush 1218175, corrupt 0, gen 0

Lots of errors on one of the cache devices, these are read/write errors, i.e. hardware problem.

 

Likely because of those filesystem has issues, run a scrub and check that there are no uncorrected errors, but note that data in any NOCOW shares can't be checked/fixed.

[JorgeB]

Didn't' check the complete syslog, more errors now on another cache device:

 

Nov 12 11:20:38 Vengeance kernel: BTRFS error (device sdb1): bdev /dev/sdh1 errs: wr 394525, rd 74047, flush 13805, corrupt 0, gen 0

Replace cables, if errors persist it can be the devices.

[Shin Zearim]

Thank you for the reply I will run a scrub and see what comes up. I also wanted to ask what could be sdh1. I only have 7 devices which includes the flash drive. 

[JorgeB]

Cache pool is indicating there's a missing device:

 

Nov 11 22:27:00 Vengeance emhttpd: cache TotDevices: 4
Nov 11 22:27:00 Vengeance emhttpd: cache NumDevices: 3
Nov 11 22:27:00 Vengeance emhttpd: cache NumFound: 3
Nov 11 22:27:00 Vengeance emhttpd: cache NumMissing: 1
Nov 11 22:27:00 Vengeance emhttpd: cache NumMisplaced: 0
Nov 11 22:27:00 Vengeance emhttpd: cache NumExtra: 0
Nov 11 22:27:00 Vengeance emhttpd: cache LuksState: 0

Likely one of them dropped offline then came back up with a different identifier, this can confuse btrfs, most commonly with SSD these issues are cable related.

[Shin Zearim]

The scrub looks like it came back fine, no uncorrected errors. I'll replace cables and post an update once that is done. Thank you for your help.