🚀 go-pugleaf

Can someone please help, I've been living with this problem for 8 months
now and am at my wits end.

Here it is:

We are using a uVax 3600 running Ultrix 2.2 with patches as a NFS server
to a network of diskfull Sun 3/60s running 4.0 (ie. all the user files
reside on the Vax and the Suns NFS mount them).

The problem is that the Vax gets a lot of "NFS stale file handle" messages
coming across its console from the Suns on the network.  These messages
can be coming across so rapidly the they bring the Vax down to its knees.
The errors can come in short spurts or a continous stream.  When they come
in a continous stream I have to go and reboot the offending Sun before
they will stop.  I've even had Suns that are logged out beat the Vax to
death and if it happens when I'm not around to find the Sun and reboot it
it will eventually crash the Vax due to filling up errlogs and running out
of mbuf space.	

I have not been able to isolate any certain process or sequence that the
sun's are doing and the problem is intermitent.

Sun and Dec have been told of the problem but untill the recent
'connectathon' did not acknowledge it as anything more than a site
specific problem.  They seem to be working on it but in the meantime I
still have to live with it and it's not doing my reputation any good.

So I thought I would throw it out to the net to see if I could get any help.

Here is an example error message on the console:

NFS server: stale file handle fs(9,14) file 237806 gen 28
             local gen 37 nlink 0, read, client address = 128.251.30.31
NFS server: stale file handle fs(9,14) file 237806 gen 28
             local gen 37 nlink 0, getattr, client address = 128.251.30.31
                           .
                           .
                           .

Any help would be appreciated.

Dennis Fail
Rockwell International
Richardson, Texas
214-996-2471

{uunet | killer}!fozzy!fail