hello, 

 

just looking for some guidance on a course of action with my corrupted BTRFS cache pool since I woke up to a wall of errors in my log. 

 

This is a relatively fresh build but has been running for 17 days flawlessly so this came out of the blue. The errors are listed below and full diagnostics is attached. It should captured the first instance of the error as the server wasn't rebooted since. 

 

I ran a full scrub but all the errors came back as unrecoverable. I did not check the FS integrity. 

 

should I just replace the drive? 

 

Apr  5 06:45:59 Tower kernel: btrfs_end_super_write: 11 callbacks suppressed
Apr  5 06:45:59 Tower kernel: BTRFS warning (device nvme0n1p1): lost page write due to IO error on /dev/nvme1n1p1 (-5)
Apr  5 06:45:59 Tower kernel: BTRFS error (device nvme0n1p1): error writing primary super block to device 2
Apr  5 06:46:00 Tower kernel: btrfs_dev_stat_inc_and_print: 3 callbacks suppressed
Apr  5 06:46:00 Tower kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 127968576, rd 52831663, flush 907641, corrupt 3, gen 0

 

 

tower-diagnostics-20250405-0644.zip

so just looking at the BRFS config log, the drive actually shows as missing: 

 

Total devices 2 FS bytes used 200.12GiB
    devid    1 size 953.87GiB used 229.06GiB path /dev/nvme0n1p1
    devid    2 size 0 used 0 path /dev/nvme1n1p1 MISSING

 

I can however see the drive as healthy in the GUI so I wonder if I should just reboot it and see if it comes back. Again, not sure what my next step should be.

ok, so I replaced the drive that was reported as MISSING in the log and the Cache is rebuilding now. The Array started normally. 

 

it's been a bit of a monologue here but all in all, I guess I'm good. 

The syslog rotated, but looks like the NVMe device was dropping offline, and most often that's not a device problem, there are some kernel options that can help sometimes, but it would be good to see the syslog showing the drop.