Package: resource-agents Version: 1:3.9.2-5+deb7u2 When using pacemaker/corosync/resource-agents to operate pgsql active/standby database cluster on debian wheezy, following error is logged regulary in /var/log/corosync/corosync.log :
Oct 25 16:38:53 pgsql1 lrmd: [2037]: info: RA output: (pgsqld:monitor:stderr) /usr/lib/ocf/resource.d//heartbeat/pgsql: 647: [: Oct 25 16:39:03 pgsql1 lrmd: [2037]: info: RA output: (pgsqld:monitor:stderr) /usr/lib/ocf/resource.d//heartbeat/pgsql: 647: [: Oct 25 16:39:13 pgsql1 lrmd: [2037]: info: RA output: (pgsqld:monitor:stderr) /usr/lib/ocf/resource.d//heartbeat/pgsql: 647: [: Oct 25 16:39:24 pgsql1 lrmd: [2037]: info: RA output: (pgsqld:monitor:stderr) /usr/lib/ocf/resource.d//heartbeat/pgsql: 647: [: Oct 25 16:39:34 pgsql1 lrmd: [2037]: info: RA output: (pgsqld:monitor:stderr) /usr/lib/ocf/resource.d//heartbeat/pgsql: 647: [: ... It is caused by using "==" in /usr/lib/ocf/resource.d/heartbeat/pgsql:647 instead of "=": [ "$1" == "validate-all" ] && exit $rc This renders pgsql RA somehow failing. It is already fixed in 1:3.9.3+git20121009-3.1 I suggest fixing "==" -> "=" for wheezy, so other RAs are not affected.
