BTRFS critical (device dm-1): corrupt leaf Error

[Ustrombase]

Hi,

I am getting issues with my server. I have a `BTRFS critical (device dm-1): corrupt leaf`. My diagnostics are attached. I researched this and I see a solution is to reformat the drive that is having the issue. I have 2 cache pools 1 for VMS and Docker and the other is for cache. I have my data backed up. However, i'm hoping to seek some help on the root cause.

edith-diagnostics-20230725-1511.zip

[Ustrombase]

I have done more research on this and RAM is coming up. I did change RAM last month, but I did a memtest recently 1 full pass with 0 issues on the RAM. It is a 64gb so it takes a very long time to complete 1 pass.

 

I could run another memtest to make it run for 2 passes, but I rather see if it is another issue first given the 1 pass of memtest recently.

[Ustrombase]

BTW I forgot to mention my server is on atm so I can still navigate to it via webUI

[Ustrombase]

Ok so I have done 6 passes of memtest with no errors so I officially don't think it's the ram. Any other ideas? I may try to recreate the docker image. 

[itimpi]

11 hours ago, Ustrombase said:

Ok so I have done 6 passes of memtest with no errors so I officially don't think it's the ram.

Just for reference passing memtest is not definitive, whereas failing it is.

[JorgeB]

write time tree block corruption detected

 

A reboot should fix it for now, but yes, this is usually a sign of a RAM issue.

[Ustrombase]

On 7/28/2023 at 4:56 AM, itimpi said:

Just for reference passing memtest is not definitive, whereas failing it is.

Yea well that is just great then...

[Ustrombase]

On 7/31/2023 at 5:30 AM, JorgeB said:

write time tree block corruption detected

 

A reboot should fix it for now, but yes, this is usually a sign of a RAM issue.

So are there any further steps to help identify root cause? I haven't had this issue again but it is frustrating to have this issue seemingly unresolved and perhaps just hibernating. To put it in another way is there a way to do a process of elimination?

[JorgeB]

There's a small possibility that it's a kernel issue, more users than usual are seeing this error after updating to v6.12, but still a very small quantity, you could convert the pool to zfs, if you also have issues with that one then it would point to a hardware problem.