I received this unRAID error notification:

unRAID file corruption: 30-07-2023 04:50
Notice [TOWER] - bunker verify command
Found 1 file with BLAKE3 hash key corruption

 

No other information was provided.  No information of which file or which drive.  No information on what to do next.  I now have to do my own research into what's going on, at a stressful time of presumed data corruption.

 

The Main tab doesn't show any errors:

 

Fortunately I was using btrfs, so I can do a scrub to look for corruption.

But I have no idea what "csum=2" means.

 

# btrfs dev stats /mnt/disk1
[/dev/md1p1].write_io_errs    0
[/dev/md1p1].read_io_errs     0
[/dev/md1p1].flush_io_errs    0
[/dev/md1p1].corruption_errs  11
[/dev/md1p1].generation_errs  0

 

I did another btrfs scrub, and now there's no errors:

# btrfs scrub start -B /mnt/disk1
scrub done for 28221102-46bb-4fa0-8602-e436c1cab5f6
Scrub started:    Sun Jul 30 16:01:10 2023
Status:           finished
Duration:         3:26:52
Total to scrub:   920.57GiB
Rate:             64.84MiB/s
Error summary:    no errors found

 

 

I also looked into my log.

It doesn't tell me what that "1 corruption" is.

I don't know if the entire drive's data is rubbish.

I don't know if the drive needs to be replaced.

Jul 30 04:50:39 Tower bunker: verified 10677 files from /mnt/disk1. Found: 0 mismatches, 1 corruption. Duration: 04:50:38. Average speed: 47.1 MB/s
Jul 30 10:30:52 Tower kernel: BTRFS warning (device md1p1): checksum verify failed on logical 1053360128 mirror 1 wanted 0x8cb7f8d8 found 0x9c0da989 level 0
Jul 30 10:30:53 Tower emhttpd: read SMART /dev/sdh
Jul 30 10:31:04 Tower kernel: BTRFS info (device md1p1): read error corrected: ino 0 off 1053360128 (dev /dev/md1p1 sector 2073728)
Jul 30 10:31:04 Tower kernel: BTRFS info (device md1p1): read error corrected: ino 0 off 1053364224 (dev /dev/md1p1 sector 2073736)
Jul 30 10:31:04 Tower kernel: BTRFS info (device md1p1): read error corrected: ino 0 off 1053368320 (dev /dev/md1p1 sector 2073744)
Jul 30 10:31:04 Tower kernel: BTRFS info (device md1p1): read error corrected: ino 0 off 1053372416 (dev /dev/md1p1 sector 2073752)

 

 

bunker -v -b3 "/mnt/user/media/Movies/[movie_name]"/
Finished - verified 1 file, skipped 0 files. Found: 0 mismatches, 0 corruptions. Duration: 00:04:56. Average speed: 18.0 MB/s

 

I haven't done anything else, but appears the error disappeared, and I don't have specific information on a file that is corrupt to initiate a restore from backup.  However, my data quality confidence is low, and may be considering a complete move to a technology that has integrated data integrity; most likely ZFS.

 

 

A sporadic error suggests RAM problems, suggest running memtest.

 

Also if the error was still there the btrfs scrub you list the affected file(s) in the syslog.