> Thanks Jeff for the detailed explanation. You mentioned delayed changelog may > have prevented this issue. Can you please tell me how to enable it? It's on by default (cluster.post-op-delay-secs in the CLI) with a value of 1. However, it's inoperative in some cases, such as when multiple fds are open for the same file. It might also be possible that a variant of the same problem could still occur e.g. if one or more bricks paused for more than the post-op delay before responding. Unfortunately, the AFR code changed a great deal recently, so I don't feel as confident in my knowledge at that level of detail as I once did. Maybe one of the AFR authors/maintainers will grace us with a more authoritative answer. _______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://supercolony.gluster.org/mailman/listinfo/gluster-users
