Problem resolved, rgmanager seems to rely on the 'status' action rather than on the 'monitor' one.
Which doesn't seem to conform to the opencf API ( http://www.opencf.org/cgi-bin/viewcvs.cgi/*checkout*/specs/ra/resource-agent-api.txt?rev=1.10) chapter 3.4.3 (is this doc outdated ?) that states that 'start' 'stop' 'monitor' and 'meta-data' are mandatory. Regards 2009/8/10, brem belguebli <brem.belgue...@gmail.com>: > > Hi, > > I have written a custom resource script that starts, stops and "meta-data" > 's with the correct return value (no error). > > Everything seems to work fine except the fact that rgmanager never runs the > custom monitor action defined in my resource metadata. > > I have lowered the interval value for monitor action to 10 ( are the > interval and timeout units in seconds by default ?) but nothing happens. > > I have added a trace to the monitor case of my resource script that ocf_log > debug "monitor started at `date +%H:%M:%S`" but it never gets there: > > > ----------------------------------------------8<-------------------------------------- > > case $1 in > > start) .................... > > ;; > > stop) .......................... > > ;; > > monitor) ocf_log debug "at `/bin/date +%H:%M:%S`, launching > monitor" > check_conf > check_status > check_md_state > if [ $? -eq 33 ]; then > repair > fi > ;; > > meta-data) cat `echo $0 | sed 's/^\(.*\)\.sh$/\1.metadata/'` > exit $OCF_SUCCESS > ;; > > *) echo "usage: $0 {start|stop|monitor|meta-data}" > ;; > esac > > ---------------------------------------->8------------------------------------------------ > > Any idea ? Thanks > > attached my meta-data > > > > > > > > > > >
-- Linux-cluster mailing list Linux-cluster@redhat.com https://www.redhat.com/mailman/listinfo/linux-cluster