I noticed my Syncthing docker was not synchronizing files and discovered I/O errors in the logs. This led me to the system log where I found this...

Jul 20 21:34:58 MEESEEKS kernel: BTRFS error (device md5p1): bad tree block start, mirror 1 want 5801970647040 have 0
Jul 20 21:34:59 MEESEEKS kernel: BTRFS error (device md5p1: state A): Transaction aborted (error -5)
Jul 20 21:34:59 MEESEEKS kernel: BTRFS: error (device md5p1: state A) in __btrfs_update_delayed_inode:1067: errno=-5 IO failure
Jul 20 21:35:02 MEESEEKS kernel: BTRFS error (device md5p1: state EA): bad tree block start, mirror 1 want 5801970663424 have 0

On the Main tab, Disk #5 is not showing any errors. But when I click on Disk 5 I can see over 8000 corruption errors on the pool device status. Running an extended SMART test resulted in no errors. How can I fix this file system issue? Thanks

meeseeks-diagnostics-20250721-1649.zip

With btrfs it's not typically easy to repair a filesystem; recommend backing up and reformatting, but before doing that, run memtest and also correct the RAM installation. Both DIMMs are on the same memory channel; this is not optimal for performance and stability.

Thanks Jorge,

By reformatting, do you mean remove the disk and re-add or replace? Parity is still valid and should rebuild the data if I understand correctly. Thanks

Edited July 22, 2025Jul 22 by meyerj31

No, parity cannot be relied upon for such a scenario. Would need to copy everything off this drive, reformat it (while in the array), and copy everything back on. No removal/replacement.

Okay please verify this action plan

1. Connect a spare hard disk via USB or SATA

2. Use Krusader to copy the contents of disk 5 to the spare disk.

3. Stop array and format disk

4. Copy contents back

5. Start array

6. Run a parity check.

Does that sound right?

Sounds good.