[Pacemaker] active-active setup with crm clone and load balancing

[Tomas Kouba]

Hello all,

I am running a simple information system that does not need to have
a backend storage.
I would like to run it in two instances on two nodes and have them in
high available and load balancing setup. So the behaviour would be
that both nodes share one IP address and requests are load balanced.
I can achieve this with configuration number one (see below).

The second part of the behavior would be that if one node dies
all requests would be served by the one node that survived.
I can achieve this with configuration number two.

My problem is that I cannot achieve both steps in one configuration.

I have followed the apache setup described in the Cluster_from_Scratch.pdf
but I must have missed some step. Configuration number one looks like this:

# crm configure show
node sbdii1.farm.particle.cz
node sbdii2.farm.particle.cz
primitive sbdii12ip ocf:heartbeat:IPaddr2 \
        params ip="147.231.25.43" cidr_netmask="24"
clusterip_hash="sourceip" \
        op monitor interval="10s"
clone sbdiiIP sbdii12ip \
        meta globally-unique="true" clone-max="2" clone-node-max="1"
property $id="cib-bootstrap-options" \
        dc-version="1.0.8-9881a7350d6182bae9e8e557cf20a3cc5dac3ee7" \
        cluster-infrastructure="openais" \
        expected-quorum-votes="2" \
        stonith-enabled="false" \
        no-quorum-policy="ignore"


Configuration number two:

# crm configure show
node sbdii1.farm.particle.cz
node sbdii2.farm.particle.cz
primitive sbdii12ip ocf:heartbeat:IPaddr2 \
        params ip="147.231.25.43" cidr_netmask="24"
clusterip_hash="sourceip" \
        op monitor interval="10s"
clone sbdiiIP sbdii12ip \
        meta globally-unique="true" clone-max="2" clone-node-max="2"
property $id="cib-bootstrap-options" \
        dc-version="1.0.8-9881a7350d6182bae9e8e557cf20a3cc5dac3ee7" \
        cluster-infrastructure="openais" \
        expected-quorum-votes="2" \
        stonith-enabled="false" \
        no-quorum-policy="ignore"


Thank you for any help,

--
Tomas Kouba