Can you create a bug for this please? Be sure to include a hb_report archive covering the period when you used the lsb script to restart the resource.
On Thu, Dec 11, 2008 at 18:38, Arndt Roth <[email protected]> wrote: > Hi *, > > > > I found a puzzling behaviour using ldirectord as a clone resource in a 3 > node setup. > > On one of the nodes "server1" (not the master) I restarted ldirectord > by hand using the lsb-script, though I know that all resources are > managed by OCF-classes. > > > > The master still displays the clone-resource correctly. > > > > Clone Set: clone_ldirectord: > > resource_ldirectord:0 (ocf::heartbeat:ldirectord): Started > server1 > > resource_ldirectord:1 (ocf::heartbeat:ldirectord): Started > server3 > > resource_ldirectord:2 (ocf::heartbeat:ldirectord): Started > server2 > > > > The other 2 nodes don't: > > > > Clone Set: clone_ldirectord > > resource_ldirectord:0 (lsb:ldirectord): Started server1 > > resource_ldirectord:1 (lsb:ldirectord): Started server2 > > resource_ldirectord:2 (lsb:ldirectord): Started server3 > > > > On all 3 nodes a config dump shows me that a third of the lrm_resources > are "lsb" now: > > > > cibadmin -Q | grep ldirec | egrep -i "ocf|lsb" > > <primitive id="resource_ldirectord" class="ocf" > type="ldirectord" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:0" type="ldirectord" > class="lsb" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:1" type="ldirectord" > class="ocf" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:2" type="ldirectord" > class="ocf" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:0" type="ldirectord" > class="lsb" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:1" type="ldirectord" > class="lsb" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:2" type="ldirectord" > class="ocf" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:0" type="ldirectord" > class="ocf" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:1" type="ldirectord" > class="ocf" provider="heartbeat"> > > <lrm_resource id="resource_ldirectord:2" type="ldirectord" > class="ocf" provider="heartbeat"> > > > > What did I misconfigure in the clone-definition? > > How comes manual process interaction changes configuration in the > cluster? > > How can I revert this or clear lrm_resources? > > > > Thanks for any help, > > > > Arndt Roth > > > > > > > > > > > > > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
