I have asked before but I think the question got lost in some other problems we had, so I will try this again. We have a cluster of machines running gluster 2.0.2 and we wish to upgrade to 2.0.9... when we tried this before, it seemed like almost all of the files were not showing up in the dir. We then realized that there is some kind of hash table difference or something going on between revisions. We were not able to make this go away by touching or doing an "ls" on the drive from the server or the client. I have seen this before but we were not sure what we did to fix it before (when we went from 1.3 to 1.4). Someone mentioned running the backend-xattr-sanitize.sh in the extras directory of the tarball, but we were talking about a production server and were not sure if that would be advisable (as if we could not access the data, we would really be in trouble). Any ideas or help would be appreciated. Thanks.