Hi All,

I have a little problem here wherin i can use your help..

1. Solaris systems in the one of the offices office can NFS mount the NetApp Filer filer1 and work properly. No problems are seen.

2. Irix machines cannot. The can mount exports from filer1 and traverse directory trees, i.e: cd unix/tools/local, etc. Thay can do an ls or 'ls -l' or many directories, but cannot do an 'ls' or 'ls -l' in others. The difference is believed to be the the directories posing problems contain symbolic links. This has not been completely demonstrated. This happens both when the Irix clients are forced to use NFSv2 via the mount option "vers=2" and when they are allowed to default to NFSv3.

3. These same Irix machine can mount a Other offices NetApp filer and have no problems. The symptoms descibed above are not seen. In turn, a other offices Irix system (there are about 500 Irix systems who are active and happy NetApp clients in other office) will see the same same probems when mounting filer1 accross the WAN.

4. The problem between Irix and NetApp centers account the filer1.

There are two observed differences betwenn the filer1 and a other filer3,

4.1 Other office is runnig NetApp Release 5.3.5R2P2 while here the filer1 is running NetApp Release 6.0.3

4.2 Other office has "nfs.tcp.enable off" while here filer1 has it set to "nfs.tcp.enable on".

5. We set "nfs.tcp.enable off" in Munich and the Solaris clients hung (they evidently preferred to use TCP over UDP). A reboot of Filer1 and the Solaris systems did not fix them. We turned NFS over TCP back on.

6. We then forced a single Irix NFS mount to use NFSv3 and use TCP via the client mount option "proto=tcp". It worked.

7. Does this symtom of "UDP not working" match a common NetApp misconfiguation?

8. Are there any know issues with UPD under NetApp Release 6.0.3?