I'm looking for a way to have a master/slave fail-over that share 2 common virtual-IPs (on Ubuntu):
- vip1: 10.0.0.171
- vip2: 10.0.0.172
- master: 10.0.0.181
- slave: 10.0.0.182
I need the following functionality:
- If both master & slave machines are up, master gets both VIPs
- If master goes down, slave gets both VIPs
- If master goes back up, it gets back both VIPs
- If only master is up, well, master gets the VIPs (of course.)
I was able to achieve this easily with heartbeat, but it seems like corosync+pacemaker is "the future." Unfortunately, finding good, concise documentation has proven difficult for me to find...
This is my heartbeat setup:
/etc/hosts:
10.0.0.181 master-host
10.0.0.182 slave-host
/etc/ha.d/ha.cf:
deadtime 3
warntime 5
initdead 10
bcast eth0
auto_failback on
node master-host
node slave-host
use_logd yes
/etc/ha.d/haresources:
master-host IPaddr::10.0.0.171/24/eth0 IPaddr::10.0.0.172/24/eth0
As a bonus (something that heartbeat seems to be unable to do on its own) it would be nice to be able to:
- Have both the master & slave be unaware of one another (i.e. they don't know each other IPs)
- they can share a password
- each knows if it's a master or a slave
Extra bonus:
- being able to add a second slave, such as slave: 10.0.0.183
Is this something that can be accomplished with corosync & pacemaker?
0 Answers