Hey. I've the following on a btrfs that's basically the system fs for my notebook: When booting from a USB stick with: # uname -a Linux heisenberg 4.17.0-3-amd64 #1 SMP Debian 4.17.17-1 (2018-08-18) x86_64 GNU/Linux # btrfs --version btrfs-progs v4.17 ... a lowmem mode fsck gives no error: # btrfs check --mode=lowmem /dev/mapper/system ; echo $? Checking filesystem on /dev/mapper/system UUID: 6050ca10-e778-4d08-80e7-6d27b9c89b3c checking extents checking free space cache checking fs roots ERROR: errors found in fs roots found 495910952960 bytes used, error(s) found total csum bytes: 481840472 total tree bytes: 2388819968 total fs tree bytes: 1651097600 total extent tree bytes: 161841152 btree space waste bytes: 446707102 file data blocks allocated: 6651878428672 referenced 542320984064 1 ... while a normal mode fsck doesn't give one: # btrfs check /dev/mapper/system ; echo $? Checking filesystem on /dev/mapper/system UUID: 6050ca10-e778-4d08-80e7-6d27b9c89b3c checking extents checking free space cache checking fs roots checking only csum items (without verifying data) checking root refs found 495910952960 bytes used, no error found total csum bytes: 481840472 total tree bytes: 2388819968 total fs tree bytes: 1651097600 total extent tree bytes: 161841152 btree space waste bytes: 446707102 file data blocks allocated: 6651878428672 referenced 542320984064 0 There were no unusual kernel log messages. Back in the normal system (no longer USB)... I spottet this: Aug 30 18:31:29 heisenberg kernel: BTRFS info (device dm-0): the free space cache file (22570598400) is invalid, skip it but not sure whether it's related (probably not)... and I haven't seen such a free space cache file issue (or any other btrfs errors) in a long while (I usually watch my kernel log once after booting has finished). Any ideas? Perhaps it's just yet another lowmem false positive... anything I can help in debugging this? Apart from this I haven't noticed any corruptions recently... just about to make a full backup of the fs (or better said a rw snapshot of the most of the data) with tar, so most data will be read soon at least once... an I would probably notice any further errors that are otherwise silent. Cheers, Chris.
