[Pacemaker] colocation rule not obeyed

[Andrew Beekhof]

On Mon, Nov 29, 2010 at 6:55 AM, Patrick H. <pacemaker at feystorm.net> wrote:
> I've got a replicated mysql with vip setup i've been trying to create, and
> it doesnt seem to be obeying my colocation rule much.
>
> Situation:
> I had shut down corosync on all (3) nodes and verified all associated
> services were off.
> I then started it up an all 3 nodes.
> When it came back up, the VIP was running on node1, while the master mysql
> service was running on node 2.
>
> # crm status
> ============
> Last updated: Mon Nov 29 05:51:06 2010
> Stack: openais
> Current DC: devms01 - partition with quorum
> Version: 1.0.9-89bd754939df5150de7cd76835f98fe90851b677
> 3 Nodes configured, 3 expected votes
> 2 Resources configured.
> ============
>
> Online: [ devms01 devms02 devms03 ]
>
>  vip    (ocf::heartbeat:IPaddr2):       Started devms01
>  Master/Slave Set: ms_mysql
>      Masters: [ devms02 ]
>      Slaves: [ devms01 devms03 ]
>
>
> # crm configure show
> node devms01 \
>         attributes standby="off"
> node devms02 \
>         attributes standby="off"
> node devms03
> primitive mysql ocf:etc:mysql \
>         params master_host="165.212.101.96" user="mailsafe_rep"
> pass="replicate" \
>         op monitor interval="10s" role="Master" \
>         op monitor interval="60s" role="Slave"
> primitive vip ocf:heartbeat:IPaddr2 \
>         params nic="eth0" iflabel="0" ip="165.212.101.96" cidr_netmask="24"
> broadcast="165.212.101.255" \
>         meta target-role="Started" is-managed="true"
> ms ms_mysql mysql \
>         meta target-role="Master" is-managed="true"
> colocation co_mysql inf: ms_mysql:Master vip
> property $id="cib-bootstrap-options" \
>         dc-version="1.0.9-89bd754939df5150de7cd76835f98fe90851b677" \
>         cluster-infrastructure="openais" \
>         expected-quorum-votes="3" \
>         stonith-enabled="false" \
>         no-quorum-policy="ignore" \
>         default-resource-stickiness="INFINITY" \
>         last-lrm-refresh="1291008843"
>
>
>
> I dont get it. Why did it decide to just ignore the co_mysql rule?


No idea, cibadmin -Q with the cluster in this state?