[SOLVED] Invoking Mover results in BTRFS csum errors

[filk]

Hi all,

 

As the title suggests, after running Mover, I am immediately presented with csum errors, as per below:

Aug 16 15:15:23 gunnhild emhttpd: shcmd (313): /usr/local/sbin/mover &> /dev/null &
Aug 16 15:15:26 gunnhild kernel: BTRFS warning (device dm-3): csum failed root 5 ino 293582 off 757051392 csum 0x0d54183c expected csum 0xb4018745 mirror 1
Aug 16 15:15:26 gunnhild kernel: BTRFS warning (device dm-3): csum failed root 5 ino 293582 off 757051392 csum 0x0d54183c expected csum 0xb4018745 mirror 1
Aug 16 15:15:26 gunnhild kernel: BTRFS warning (device dm-3): csum failed root 5 ino 293582 off 757051392 csum 0x0d54183c expected csum 0xb4018745 mirror 1
Aug 16 15:15:26 gunnhild kernel: BTRFS warning (device dm-3): csum failed root 5 ino 293582 off 757051392 csum 0x0d54183c expected csum 0xb4018745 mirror 1
Aug 16 15:15:31 gunnhild kernel: BTRFS warning (device dm-3): csum failed root 5 ino 354615 off 1305513984 csum 0x73d2e631 expected csum 0x8ae53185 mirror 1
Aug 16 15:15:31 gunnhild kernel: BTRFS warning (device dm-3): csum failed root 5 ino 354615 off 1305513984 csum 0x73d2e631 expected csum 0x8ae53185 mirror 1
Aug 16 15:15:31 gunnhild kernel: BTRFS warning (device dm-3): csum failed root 5 ino 354615 off 1305513984 csum 0x73d2e631 expected csum 0x8ae53185 mirror 1
Aug 16 15:15:31 gunnhild kernel: BTRFS warning (device dm-3): csum failed root 5 ino 354615 off 1305513984 csum 0x73d2e631 expected csum 0x8ae53185 mirror 1

After researching the forum, I had suspected this was down to faulty RAM or cache drive at first. A memtest corroborated this theory and I thought it was the root cause.

 

I replaced the RAM and ran another memtest which came back with 0 errors after 14 passes. The cache drive was old and low capacity so this also was changed to an MX500.

 

However, I'm still getting the same BTRFS errors and now I'm at a loose end as to what to do.

 

FWIW, Mover does infact move the files from the Cache onto the Array as expected. It's not like it ever fails.

 

A side note as I'm not sure if it's related or not: I also received the error below. It's worth mentioning it could be because the box was off for around 2 weeks and potentially could be a result of that?

Aug 16 15:19:46 gunnhild ntpd[1675]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized

 

Any help would be greatly appreciated!

 

Edited August 16, 2020 by filk
spelling

[JorgeB]

15 hours ago, filk said:

However, I'm still getting the same BTRFS errors and now I'm at a loose end as to what to do.

Run a btrfs scrub on that device, it will list the corrupt files on the syslog, those files need to be deleted/replaced.

[filk]

Thank you, @johnnie.black.

 

Removing the referenced files in the scrub seems to have resolved the issue.