Hi, Using glusterfs 3.2.1 in a two node setup with NFS Clients (some of them are too old, to support a gluster-fuse client). Some files are moved at a client, but the files don't get moved. What I see then is, that the target folder got only created on one of the backends, but not on the other. Removing an recreation of this folder doesn't help. Other files changes and folder-creations got replicated. Nothing relevant shows up in the logs. Only when chowning the directory, I get some AFR messages. Solution: Hard-Killing all gluster-prozesses on booth servers and restarting glusterfs. Now it works again. But it's rather boring and frustrating to keep restarting the gluster-prozesses every day or so... Anybody with any good ideas whats happening here? Why would replication stop on a handfull of folders, but continue to work for others? Any advise would be appreciated! Regards, Georg
