I recently installed heartbeat on Redhat, after I started the heartbeat service, it bring on all services I defined in haresware, the content in this file like node1 10.3.254.111 httpd sendmail dhcpd. After I started heartbeat, I saw the IP address for eth0:0 is 10.3.254.111. The IP address for eth0 is 10.3.254.112, I can ping 10.3.254.112 from outsite, but I cann't ping 10.3.254.111 from outsite. Default gw for this machine is 10.3.254.1, maybe I need specify the netmask and broadcast on resource file more detail like 10.3.254.111/24/eth0, because I assign the 10.3.254.112/24 to the eth0. Is this correct?

Hello,

I am not familiar with the heartbeat service. I do not know why you would want to ping both the physical and virtual adapters externally if on the same subnet. Is that what you want here?

I would suggest you try the following.

1. Confirm if you can ping from the outside with heartbeat stopped.

service heatbeat stop
(remote) ping again

2. Check that IPTables is not getting in the way. The rules may need to be changed for the VLAN. The "Masquerade" rule particularly is common and comes straight to mind here.

service iptables stop

3. I have played with two DSL connections, before and found scenarios where a ping packet comes in one interface, and goes out the other one. (A routing issue)

A good thing to do here is use tcpdump. Use it on your PC that does not reply, and see if you are getting the requests. You may find that eth0 has the request, and eth0:0 has the reply. (or vice versa)

tcpdump -i eth0

If this is the problem, you may need to play with your routing rules.

If you are still confused, use a 3rd PC to monitor the LAN. Use a dumb hub if you can get one.

If things are still not working, post some replies from tcpdump.