Everything works fine for a while after reboot. After some amount of time (less than a couple of hours), networking stops working. Can't ping any remote IP addresses by name or address. Other systems on network (e.g., Solaris, NT., etc. are fine). When problem happens, "netstat -in" output, ifconfig output, all looks okay. If I do "ifconfig eth0 down" and then "ifconfig eth0 up", then networking starts working again.
Is this on a machine with an AMD processor? If so, it could be connected with one of my systems with an AMD K6-2/350 processor that also locks up after a couple of hours.
See my bug #6258. I have a number of machines locking up. Thought it was the NICS, replace em but it still happens. I suspect SMP now and am running stock 2.2.12. I have reports that ifconfig eth0 down up restored connectivity.
Sorry, it was bug #9258.
I have a similar problem on two identical machines on different networks, both stock redhat 6.1 installs with p-III and 3c509c cards. They stop talking through the gateway but can talk to machines on the local network. If left alone they will start talking outbound after a random amount of time (10-60 minutes or so). If I drop the default route and add it back they start talking to the gateway. I'm not sure if it's the kernel since I have 6.0 machines that I've upgraded to 2.2.12 and don't have that problem.