NFS Problems in 2 Subnetworks

HugOnline · 04-08-2003, 06:43 AM

Hello,

Computer X is connected to subnetA and subnetB via 2 eth-interfaces. Computer Y is connected to subnetB and serves NFS. These NFS can be mounted by other computers in subnetB but not by computer X, which can only connect to NFS in subnetA and this is the problem. Exports is right (open for all), portmapper is working (rpc.info -p is ok),ping is working, in /var/log/messages of Computer Y mount is allowed, but computer X always gets signal : mount: RCP: time out

Where is the mistake ?

Regards

HugOnline

bentz · 04-08-2003, 09:15 AM

The problem must be with computer X. Most likely it's a network configuration issue because if it only had one interface, it'd be working like all your other NFS clients. Take a look at the 'route' output on X, and see if the machine is aware of a path to get to machine Y. Can you use any other services on machine Y, from machine X? Can you ping Y from X?

HugOnline · 04-09-2003, 01:41 AM

Thanks for your response !!!
(Ping works and route is ok as well)

I found the mistake, but not really: I stopped the iptables and then it worked - but if it was an security problem before, why could I reach the other computers in SubnetA with NFS ? I didn't find the switch to force iptables just to eth1 and not eth0 .

Regards HugOnline

bentz · 04-09-2003, 11:15 AM

Well, for example, to block incomming http requests to eth1, but not eth0, you'd use:

iptables -t filter -I INPUT -p TCP -i eth1 --dport 80 -j DROP

To view this interface specification, you'll need to use:

iptables -t filter -L INPUT -v