NFS hanging

From: Paul Dlug (paul@aps.org)
Date: Tue Jun 24 2003 - 18:06:36 EDT


We have 3 or 4 sun servers experiencing the same issue. All are
mounting NFS file systems containing user home directories off a
Network Appliance filer. The filesystem seems to "hang", when untarring
a file it will progress very very slowly, the next attempt will go very
quickly and subsequent tries will hang sporadically. Users are noticing
this behavior when opening mailboxes in pine that are located in their
home directories. They'll be working just fine, then hit one mailbox
which will wait for 30sec-1min to open. Turning on debugging in pine
shows an excessing number of "alarm_signal()" in the debug file.

nfsstat -c shows some timeouts and retries but they're not that
excessive:
Connectionless:
calls badcalls retrans badxids timeouts newcreds
2805125 146 15158 0 15153 0
badverfs timers nomem cantsend
0 9602 0 0

Load on the filer is extremely low, nothing above 2-8% cpu utilization
with only a few hundred NFS ops/sec. Network connections all seem fine
(100mbps full duplex from server to switch, 1000mbps full duplex from
filer to switch). All are on the same switch with no collisions or
network errors of any kind.

Does anyone have any suggestions for debugging this? I'm rapidly
reaching my limits.

Thanks,
Paul
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:26:38 EDT