TCPIP 5.4 ECO4 NFS Server problem with non-HP clients

[ALLANO Jean-Marie]

After upgrading our NFS server from :
- OpenVMS 7.3
- TCPIP 5.1 ECO4
To :
- OpenVMS 7.3-2
- TCPIP 5.4 ECO4

only the Tru64 unix NFS clients continue to have a good access to this server.
SUN/Solaris and IBM/AIX clients have problems
such as very slow access or no acces at all.

I have tried to tune the NFS server but with no results.

Is this a known problem ?

Thank you for your help on this

[ALLANO Jean-Marie]

I precise that i also have such messages on
the OpenVMS NFS Server during IBM NFS client access :

CMSSRV_SYSTEM>
%%%%%%%%%%% OPCOM 17-JUN-2005 14:08:44.65 %%%%%%%%%%%
Message from user NFS Server on CMSSRV
Accepted connection from host 10.200.200.195 port 39241


CMSSRV_SYSTEM>
%%%%%%%%%%% OPCOM 17-JUN-2005 14:09:01.14 %%%%%%%%%%%
Message from user NFS Server on CMSSRV
svcktcp_reply: mbuf_send returned 32

CMSSRV_SYSTEM>
%%%%%%%%%%% OPCOM 17-JUN-2005 14:09:01.14 %%%%%%%%%%%
Message from user NFS Server on CMSSRV
rfs_dispatch: sendreply failed IP address: 10.200.200.195

On the IBM NFS client, i get the result of my commande, for example the listing of a NFS directory, but after a very very long time.

[Ian Miller.]

I think the version of NFS protocol may have changed from TCPIP V5.1 to V5.4
Do you know which version Solaris and AIX are trying to use.