I am having the same problem on several systems running dhcp and pump.  This problem is that the network is up and fully functional, but after xxx amount of time eth0 disappers from ifconfig.  Review of /var/log/messages shows no indications of problems or why it shut down.  System is using dynamic ip on eth0 and static on eth1.  Eth1 always stays up.  It appears to happen after a length of time that eth0 is unused.  It might also be noted that if I shut down the firewall and flush its rules, plus return system to accept all packaets on all rules and use ifup eth0, it always fails.  If I use /etc/rc.d/init.d/network restart it usually starts back up again.

I have been unable to determine when the eth0 actually fails.  It has happened as quickly as a few minutes and has lasted without problems for as long as a few days....so I don't think it is dhcp lease time related.

Any help in finding the cause is much appreciated since the systems are all personal/business/client gateways.

I can send any other information you request from these systems (logs/installed rpms/or anything else you request).  I might add they all have different hardware configurations and eth cards.  They all have the latest updates available through up2date.  The only common denominator is the same ISP, though redhat v5.2, 6.0, 6.1, and 6.2 have all worked great previously.  I also thought maybe it was due to the upgrade from v6.x to 7.0 so I used one system as a test and did a clean install of v7.0 on a new hard drive, the problem is still present.

Thanks,
Ed Wiget