Hi,
I am using bitrot feature in 3.10.1 and I have few issues to point.
1. during a disk failure situation or some reason to run "volume start force" command triggers bitrot crawl process in all the bricks in the volume. it runs on disk which has already completed crawl process. I feel it's unnecessary load on the running cluster.
2. Restarting glusterd service in individual node also triggers bitrot crawl process in all the bricks but within the service restarted node.
3. Running "scrub status" creates a link in brick process like "/proc/brick-proc/fd/.glusterfs/quarantine", it keeps on increasing. if I run scrub status command 20 times I could see 20 links as mentioned.
4. scrub takes a long time to complete process even though it completes files hash check.
5. When file marked as bad by scrub in brick, reading file from mount point gives a low throughput.
regards
Amudhan
_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://lists.gluster.org/mailman/listinfo/gluster-users
