NFS problem: "RPC: Port mapper failure"

[Rob]

Igor Pokrovsky wrote:
> On Sun, Dec 28, 2003 at 09:20:41PM +0900, Rob wrote:
> 
>>Hi,
>>
>>I am running two FreeBSD-Stable PCs.
>>
>>One is an NFS server and the other NFS client. Everything used to
>>work fine until recently. I suspect that either the new kernel is the
>>problem (although there are no complaints on the stable mailing list)
>>or the network administration has changed some things.
> 
> You can try to check if nfsd and portmap really started on server:
> 	sockstat | grep portmap
> and
> 	sockstat | grep nfsd

This is what I get.

The NFS server is 147.46.44.183  with gateway 147.46.44.1
The NFS client is 147.47.254.184 with gateway 147.47.254.1

On the NFS server:
	$ sockstat | grep portmap
	daemon   portmap  79693    3 udp4   *:111       *:*
	daemon   portmap  79693    4 tcp4   *:111       *:*
	$ sockstat | grep nfsd
	root     nfsd        87    3 tcp4   *:2049      *:*
	$ sockstat | grep mount
	root     mountd      84    3 udp4   *:1023      *:*
	root     mountd      84    4 tcp4   *:1023      *:*

On the client:
	$ rpcinfo -p 147.46.44.183
	rpcinfo: can't contact portmapper: RPC: Remote system error - Operation timed out
	

On another computer with the same gateway as the server:
	$ rpcinfo -p 147.46.44.183
	program vers proto   port
	100000    2   tcp    111  portmapper
	100000    2   udp    111  portmapper


Any idea what could block the contact to the portmapper?
Could a network setup at one of the gateways cause these trouble, such
as a firewall (though I don't know anything about firewalls!) ?
Secure shell connections still work fine between all the computers!

Thanks for your help.
Rob.