NFS Client Performance

[Kevin C. Babbitt]

I'm running a D320. I have an NT server running the InterDrive NFS Server software. This NT server hosts many ~ 10Mbyte .tif images. The D320 (10.20) mounts the NT drive these images are on and they're used by various applications.
"Something" has happend to performance. The D320 rsize and wsize = 8192.
NFS server (NT) and client (HP-UX) are on different segments through a router.
If some images are MOVED to the D320 they render fine using the same application.
The images render fine from an NT workstation and mapped drive.
# nfsd=4, #biod=4 on the D320.
A Network General Sniffer on both segments shows excellent utilization data - doesn't seem to be a bandwidth issue.

Anybody? A helping hand?
Thanks...

[Rick Garland]

On the UNIX side, can run the nfsstat command to get some info specific to the NFS traffic.

[Kevin C. Babbitt]

The only thing I see here that is mentioned in any literature is the 'timeout'. After reseting the counters and performing a test I had 313 Client rpc calls and 35 timeouts.
Everyting else was at 0.

[Tim Malnati]

Timeout errors would suggest that the NT server is having some trouble serving the file. These are a function individual communications to get the file across instead the whole file. Is this NT box serving requests from other machines as well? If so you might be better off using something like Samba on the HPUX box instead. This way the NT machine does not have the additionl conversion load of Interdrive software.

[Kevin C. Babbitt]

I think I concur. These images are on the NT box because it also serves as our Intranet server. There are apps running on it that are the 'primary' users of the resource. Other than the intranet apps the most use is in the NT environment.

Thanks for the help. It's NT - REBOOT