|
Hi! I am running gluster 3.4.2 (latest stable)
and have with bricks setup as a replicated volume. I am simulating a disaster by wiping out
all the files on one of the brick (storage-1b) while the other one (storage-1a)
holds the complete content What I’ve done so far to simulate: # Remove storage-1b brick gluster volume remove-brick vol0 replica 1 storage-1b:/mnt/gluster-vol0 # On storage-1b, remove all files and reset
attributes rm –rf /mnt/gluster-vol0/* rm –rf /mnt/gluster-vol0/.glusterfs setfattr -x trusted.gfid /mnt/gluster-vol0 setfattr -x trusted.glusterfs.volume-id /mnt/gluster-vol0 # Add the brick back gluster volume add-brick vol0 replica 2 storage-1b:/mnt/gluster-vol0 # perform full rehal gluster volume heal vol0 full Once the reheal has completed, I’ve noticed
some of the files have different sizes. Some of them are short of few KBs, but I
notice there is one falls short of nearly 100MB. Even I try to force a full re-heal of that
particular file by removing it and together with the hex file in .glusterfs,
the result after a re-heal is the same. rm –f /mnt/gluster-vol0/file rm –f /mnt/gluster-vol0/.glusterfs/c7/81/c781e91a-4ffa-47a8-ba98-da111d77075b gluster volume heal vol0 full What could be the reason? Most of
these files are sparse files. Can that be a problem? Please advise. Thanks, Adrian |
_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://supercolony.gluster.org/mailman/listinfo/gluster-users
