[Linux-HA] Redundant Rings "Still Not There?"

[Pavlos Parissis]

On 24 October 2010 22:09, Robinson, Eric <eric.robinson at psmnv.com> wrote:

>
> >> But now could someone please elaborate on Dejan Muhamedagic's
> original
> >> comment that started the thread? What does "redundant rings are still
>
> >> not there" mean? Is a three-node cluster an unreliable setup because
> >> Corosync and/or Pacemaker are not really ready for that?
> >>
>
> >not Pacemaker - Corosync. I agree with Dejan. I also use Heartbeat for
> clusters communication layer.
>
> So you use Heartbeat+Pacemaker instead of Corosync+Pacemaker? How does
> Heartbeat handle 3-node clusters? Does it use the same "ring" metaphor
> and switches for cluster communication?
>
>
I have a 3-node cluster with heartbeat and works fine.
Here is my ha.cf to get an idea
root at node-01 ~]# cat /etc/ha.d/ha.cf
autojoin none
use_logd on
debug 0
mcast eth0 239.0.0.43 694 1 0
bcast eth1
warntime 5
deadtime 15
initdead 60
keepalive 2
node node-01
node node-02
node node-03
crm respawn
apiauth stonith-ng      uid=root

As you can see there is nothing special for building 3-node cluster.

Cheers,
Pavlos