2025-03-21T14:13:49+01:00 Tower kernel: BTRFS warning (device sdg1): csum failed root 5 ino 426679 off 5209911296 csum 0xc1e1b682 expected csum 0x7e34b248 mirror 1
2025-03-21T14:13:49+01:00 Tower kernel: BTRFS error (device sdg1): bdev /dev/sdg1 errs: wr 0, rd 0, flush 0, corrupt 1, gen 0
2025-03-21T14:13:49+01:00 Tower kernel: BTRFS warning (device sdg1): csum failed root 5 ino 426679 off 5209911296 csum 0xc1e1b682 expected csum 0x7e34b248 mirror 1
2025-03-21T14:13:49+01:00 Tower kernel: BTRFS error (device sdg1): bdev /dev/sdg1 errs: wr 0, rd 0, flush 0, corrupt 2, gen 0

 

This means btrfs is detecting data corruption, run a scrub, also a good idea to run memtest.

1 minute ago, JorgeB said:

This means btrfs is detecting data corruption, run a scrub, also a good idea to run memtest.

Thanks for the reply. Running scrubbing now, should I run Balance too?

How to run memtest on Unraid?

4 minutes ago, Vitek said:

should I run Balance too?

Not for this.

 

4 minutes ago, Vitek said:

How to run memtest on Unraid?

It's a boot menu option.

@JorgeB How can I fix this Uncorrectable error?

You will need to delete the affected file or restore from a backup, it should be listed in the syslog during the scrub.

@JorgeBafter running memtest for about 20 minutes server restart (happen twice already), does it mean I need to replace memory? Can't see any errors before restart because it happen suddenly.

Server restarting on its own is almost always hardware, but necessarily RAM, if you have multiple DIMMs try running the test with one at a time.

@JorgeB I have a time machine share and every time macOS creates a backup, scrub finds an Uncorrectable error in that share. I have tried to remove everything from the time machine share I created a new backup and again Uncorrectable error has been discovered by scrub command. Is something wrong with time machine backup or my SSD/ram?

 

 

also memory SPD shows only 2GB instead of 16

Edited March 22, 20251 yr by Vitek

Please post how the path to the corrupt file shows on the syslog.

2025-03-22T11:44:34+01:00 Tower kernel: BTRFS info (device sdg1): scrub: started on devid 1
2025-03-22T11:51:59+01:00 Tower kernel: BTRFS error (device sdg1): unable to fixup (regular) error at logical 864349323264 on dev /dev/sdg1 physical 212596424704
2025-03-22T11:51:59+01:00 Tower kernel: BTRFS warning (device sdg1): checksum error at logical 864349323264 on dev /dev/sdg1, physical 212596424704, root 5, inode 478510, offset 22249472, length 4096, links 1 (path: time machine/iMac-mac.sparsebundle/bands/413)
2025-03-22T11:56:59+01:00 Tower kernel: BTRFS info (device sdg1): scrub: finished on devid 1 with status: 0

13 minutes ago, Vitek said:

time machine/iMac-mac.sparsebundle/bands/413

And if you delete this, it still finds errors on the next scrub? And what path does it show?

No, if i delete it (or whole time machine backup share) there will be no scrubs error, but when i create new share and do a mac backup scrub will find error in that share 

Always the same file or a different one?

Random file but always in the same share (time machine backup)

 

i know i need to replace RAM because memtest shows me a few errors before server restart but could it be related or I need to replace SSD too?

Edited March 22, 20251 yr by Vitek

17 minutes ago, Vitek said:

i know i need to replace RAM because memtest shows me a few errors

Any RAM error is serious, and can cause all sorts of unpredictable behaviour.  You need to get this resolved and then see if you are still getting the issue with the time machine share.

Ok I will do scrub once I replace RAM

If you have RAM errors, you need to replace that now, you should never try to use the server with known bad RAM.