Hi folks:
I'm running heartbeat 2.1.3, kernel 2.6.18 in Centos 5. Two nodes on
this scenario:
( WAN )
||
||
( Router )
||
||
[ switch ]
||
||
-----------------------
| |
eth0 eth0
NODE1 eth2 <==> eth2 NODE2
eth1 eth1
| |
-----------------------
||
||
[ switch ]
||
||
( LAN )
I'm using virtual IPs eth0:0 and eth1:0 when a node starts acquiring
resources.
eth2 is used to join both nodes using a crossover ethernet cable.
I'm trying to improve my understanding about heartbeat communication,
failover behavior and some stuffs, so I have some questions:
1. What do I use bcast/ucast/mcast directive for? Is it suposed that
heartbeat nodes communicate each other trough interfaces asigned in
bcast/ucast/mcast?
2. What interfaces should I use in bcast/ucast/mcast? Do I need to use
more than a single interface? I think using only eth2 it would be
enough... is it?
3. What do I use ping directive for? I'm using ping pointing to the
router IP address so I can know when eth0 fails... is that right?
4. If I use ping to the router IP address to know when eth0 fails... How
can I know when eth2 fails? I know it's not recommendable to use the
other node IP address as a ping node.
5. Should I use all eth0, eth1 and eth2 with bcast/ucast/mcast
directive? I think if eth2 fails then both nodes could still communicate
trough eth0 or eth1... is that right?
5. What communication method should I use between bcast, ucast and
mcast? I was reading in mailing list archives that bcast sometimes
causes floods, delays or similar problems.
I hope someone can help me because documentation on linux-ha.org is not
very clear in some cases, and finding answers to my questions in mailing
list archives is not always easy. I'm just trying to understand better
the heartbeat behavior.
Thanks in advance
(sorry, my english is poor =\)
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
