> NFS uses the ports 2049/tcp 2049/udp, but you can obtain them > from the server using rpcinfo with: > > rpcinfo -p localhost|grep "nfs"|grep "tcp"|tr ' ' '\n'|grep -v > "nfs"|tail -n 2|head -n 1 > > rpcinfo -p localhost|grep "nfs"|grep "udp"|tr ' ' '\n'|grep -v > "nfs"|tail -n 2|head -n 1 > > and for the services quota, lock, stat and mount we use the same method > but changing the grep "nfs" by grep "mount", by example. Lockd only > uses udp, quota uses tcp and udp, mount uses tcp and udp and stat uses > also tcp and udp. > > For portmap we use the same method but changing nfs by portmapper. It > uses tcp and udp. Normally it uses 111/tcp and 111/udp. I generally don't have trouble finding the assigned ports since I have a tight ruleset and good logging. The problem occurs whenever I reboot the server or restart the service. Portmapper assigns new ports and all the client connections fail. I then have to check the log for failed connections and open new ports. At this point I have over 200 ports opened (only a couple of which are used at any given time). Jason indicates I can force the daemons to use certain ports so I'm going to look into that. Thanks for your help. Jeff
