This is the guide I followed for building an iPrint cluster using rcheartbeat:

Building an iPrint Cluster Using Heartbeat and OES Linux | Novell User Communities

Every day everyone loses connectivity to their printers because the shared IP that travels between Node A and Node B is unresponsive to anyone outside the subnet the servers reside in. To get print services working again I restart rcheartbeat on the node that is the master, which forces the IP to the other node.

I dug through some logs and found this in /var/log/messages:

Apr 13 13:09:27 prt2 syslog-ng[2047]: Changing permissions on special file /dev/xconsole
Apr 13 13:09:27 prt2 syslog-ng[2047]: Changing permissions on special file /dev/tty10
Apr 13 13:09:27 prt2 kernel: martian source 10.4.55.50 from 10.4.55.50, on dev eth0
Apr 13 13:09:27 prt2 kernel: ll header: ff:ff:ff:ff:ff:ff:*MAC-omitted*
Apr 13 13:09:27 prt2 kernel: martian source 10.4.55.50 from 10.4.55.50, on dev eth0
Apr 13 13:09:27 prt2 kernel: ll header: ff:ff:ff:ff:ff:ff:*MAC-omitted*
Apr 13 13:09:27 prt2 kernel: martian source 10.4.55.50 from 10.4.55.50, on dev eth0
Apr 13 13:09:27 prt2 kernel: ll header: ff:ff:ff:ff:ff:ff:*MAC-omitted*
Apr 13 13:09:27 prt2 kernel: martian source 10.4.55.50 from 10.4.55.50, on dev eth0
Apr 13 13:09:27 prt2 kernel: ll header: ff:ff:ff:ff:ff:ff:*MAC-omitted*


10.4.55.50 is the internal intranet IP that dictates which node is the master (all devices point to this address to print).

Anyone have ideas of whats going on?