> we tried to reproduce the freeze, but instead a classical hang showed up. This one is nice as everything seems fine with the boxes, only glusterfs mountpoints hang on both clients. There are no unusual outputs in the standard glusterfs logfiles, so we can only show you this: > These are strace output from the hanging processes, I don't know if this is really useful for you. Stephan. When you face this hang, can you send 'kill -11' to all your glusterfs/d processes and post the log files? Forcing a segfault will dump the list of pending syscalls in the logfile and will give us some place to start looking. Thanks, Avati
