https://bugzilla.novell.com/show_bug.cgi?id=203519 nfbrown@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |nfbrown@novell.com Status|NEEDINFO |ASSIGNED Info Provider|nfbrown@novell.com | ------- Comment #22 from nfbrown@novell.com 2006-09-24 20:36 MST ------- NFS over UDP is nearly as good as NFS over TCP except on very unreliable networks. It *should* work. If changing to TCP helps, it is often just hiding a problem somewhere else in the stack. The fact that you are getting "nfs: server <xxx> timed out" suggests that you are using 'soft' mounts instead of 'hard'. This is generally considerred a bad idea. With 'hard' mounts it will retry indefinitely so it might work eventually ... it might just hang forever. Knowing which might be helpful. It would be worth trying the mount option 'wsize=1024'. This was avoid sending large UDP packets and so avoid fragmentation and should improve reliability on an unreliable network. In general, to diagnose NFS problems it helps a LOT to get a tcpdump trace. i.e. on both the server and the client run tcpdump -w /tmp/trace -s 0 port 2049 and then cause the problem to happen. Then interrupt both tcpdumps and place the trace files somewhere accessable (or attach them if they aren't too big). I'm not aware of an recent NFS changes that might have caused this problem, but that doesn't mean there aren't any. Another thing that might be worth testing is ping with varying packet sizes. e.g. ping -i 0.1 -s 8192 servername see if you can find an interval or a size at which it starts dropping packets. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.