Hi, 

 

I am running a windows 11 and it's running fine and no issues but I have noticed these errors on my syslog and wondered if anyone had seen similar. The drive its complaining about is where I have the vdisk located. Its a brand new nvme wd blue, so not sure why these are alerting considering I dont see any errors elsewhere, even in grafana logging

 

kernel: BTRFS warning (device nvme1n1p1): csum failed root 5 ino 261 off 23526653952 csum 0x0f9d0bc3 expected csum 0x791bcc50 mirror 1
Dec 23 18:07:05 eljtech-core kernel: BTRFS error (device nvme1n1p1): bdev /dev/nvme1n1p1 errs: wr 0, rd 0, flush 0, corrupt 30, gen 0
Dec 23 18:08:03 eljtech-core kernel: BTRFS warning (device nvme1n1p1): csum failed root 5 ino 261 off 5651419136 csum 0xe23591e0 expected csum 0xcc7ecd01 mirror 1
Dec 23 18:08:03 eljtech-core kernel: BTRFS error (device nvme1n1p1): bdev /dev/nvme1n1p1 errs: wr 0, rd 0, flush 0, corrupt 31, gen 0

 

btrfs is detecting data corruption, run a scrub and if there are uncorrectable errors look at the syslog for the affected files.

16 hours ago, JorgeB said:

16 hours ago, JorgeB said:

btrfs is detecting data corruption, run a scrub and if there are uncorrectable errors look at the syslog for the affected files.

yeah il take a look at that and report back

 

So tried a scrub and no errors were found.

 

still getting these errors though in system log

 

Dec 24 11:06:59 eljtech-core kernel: BTRFS warning (device nvme1n1p1): csum failed root 5 ino 261 off 26375880704 csum 0x49a179cf expected csum 0x8647461a mirror 1
Dec 24 11:06:59 eljtech-core kernel: BTRFS error (device nvme1n1p1): bdev /dev/nvme1n1p1 errs: wr 0, rd 0, flush 0, corrupt 75, gen 0

 

Only seems to be when my VM is running.

 

There is no errors on the main dashboard or from my grafana dashboard related to that disk 

Edited December 24, 20223 yr by jowiedog

Suggest moving the vdisk elsewhere, then moving it back to see if it helps.

yeah will give that try cheers

I found the exact same thing with libvirt. Linux images seemed to work fine but the corruption errors were happening on my Windows VMs. Was not a disk failure as I tried on 2 NVMes and a SATA SSD.

 

The solution to this is you need to

chattr +C /path/to/qemu/images

on the directory holding the Qemu images and re-create the RAW image in there (so it has nodatacow flag set now).

 

Found this solution https://www.spinics.net/lists/linux-btrfs/msg25949.html

Edited April 18, 20251 yr by dingobear

12 hours ago, dingobear said:

The solution to this is you need to

This is not a general issue, they should still work fine with checksums enabled, there may be some other issue, or a compatibility problem.