I have a glusterfs of three servers and after a bit of troubleshooting i concluded that a file has a gfid split-brain.
**Server1**
# file: srv/glusterfs/gbrick/file1
trusted.afr.dirty=0x000000000000000000000000
trusted.afr.shared-client-1=0x000000050000000100000000
trusted.bit-rot.version=0x020000000000000059b166800006516b
trusted.gfid=0x4ca01cfe117c4115a049ae108d58271b
**Server2**
# file: srv/glusterfs/gbrick/file1
trusted.afr.dirty=0x000000000000000000000000
trusted.afr.shared-client-1=0x000000050000000100000000
trusted.bit-rot.version=0x020000000000000059b166800006516b
trusted.gfid=0xa8a1718e0bd64783b535b6ff0e26699a
**Server3**
# file: srv/glusterfs/gbrick/file1
trusted.afr.dirty=0x000000000000000000000000
trusted.afr.shared-client-1=0x000000050000000100000000
trusted.bit-rot.version=0x020000000000000059b166800006516b
trusted.gfid=0x4ca01cfe117c4115a049ae108d58271b
It looks like the trusted.gfid
on server2 is different compared to server1 and server3. Is there any way to tell glusterfs to update the file on server2 with the one of server1 (or server3) ?
from what i read at https://staged-gluster-docs.readthedocs.io/en/release3.7.0beta1/Features/heal-info-and-split-brain-resolution/
'Note that entry/gfid split-brain resolution is not supported.'
but if you find this page like me with non-gfid splitbrain scenario - these can be healed.