[Pacemaker] Problem with migration-threshold and fail-count if the cloned resource fails on both nodes

Klaus Darilion klaus.mailinglists at pernau.at
Thu Mar 3 06:03:57 EST 2011


I have node A (armani) and node B (bulgari). I have configured a virtual
IP address and Kamailio (SIP proxy) as cloned resource running on both
nodes. If Kamailio failes two times on the node with the IP address, the
IP address is migrated.

The problem happens when Kamailio fails on both nodes at the same time,
then the cloned resource gets swapped between node A and B, and the IP
address is not migrated anymore in case of Kamailio failures.

See attached log for details of the scenario. The config is below. Is
there something wrong with my config to handle this properly or is it a bug?

I think the problem is related that the clones resource is swapped
during downtime of both Kamailio resources. Is it possible to stick a
cloned resource to a certain node?


node armani \
        attributes standby="off"
node bulgari \
        attributes standby="off"
primitive failover-ip ocf:heartbeat:IPaddr \
        params ip="" \
        op monitor interval="3s"
primitive kamailio lsb:kamailio \
        meta migration-threshold="2" failure-timeout="30" \
        op monitor interval="10" timeout="10"
clone cloneKamailio kamailio
colocation colo_ip_with_kamailio inf: failover-ip cloneKamailio
property $id="cib-bootstrap-options" \
        cluster-infrastructure="openais" \
        expected-quorum-votes="2" \
        stonith-enabled="false" \
        no-quorum-policy="ignore" \
        cluster-recheck-interval="5s" \
rsc_defaults $id="rsc-options" \
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: pacemaker-log.txt
URL: <http://lists.clusterlabs.org/pipermail/pacemaker/attachments/20110303/d73005b7/attachment-0002.txt>

More information about the Pacemaker mailing list