On Wed, Nov 10, 2010 at 8:07 PM, Michael Schwartzkopff <mi...@clusterbau.com> wrote: > Hi, > > My cluster shows some strange errors. I (read: "the custumer") want to create > some strange constraints for my cluster. First of all, I want to run the DRBD > master only on a node that has a specific attribute set DRBD_Master. > > Second the filesystem should be mounted to different directories on the > different > nodes. > > So I have the following configuration: > > primitive resDRBD ocf:linbit:drbd \ > params drbd_resource="r0" > xml <primitive class="ocf" id="resFS" provider="heartbeat" type="Filesystem"> > \ > <instance_attributes id="resFS-instance_attributes-node1" score="3"> \ > <rule id="rule-FS-node1" score="INFINITY"> \ > <expression attribute="#uname" id="exp-node1" > operation="eq" > value="node1"/> \ > </rule> \ > <nvpair id="resFS-instance_attributes-device-1" name="device" > value="/dev/drbd0"/> \ > <nvpair id="resFS-instance_attributes-fstype-1" name="fstype" > value="ext4"/> \ > <nvpair id="resFS-instance_attributes-directory-1" > name="directory" > value="/mnt/dir1"/> \ > </instance_attributes> \ > <instance_attributes id="resFS-instance_attributes-node2" score="2"> \ > <rule id="rule-FS-node2" score="INFINITY"> \ > <expression attribute="#uname" id="exp-node2" > operation="eq" > value="node2"/> \ > </rule> \ > <nvpair id="resFS-instance_attributes-device-2" name="device" > value="/dev/drbd0"/> \ > <nvpair id="resFS-instance_attributes-fstype-2" name="fstype" > value="ext4"/> \ > <nvpair id="resFS-instance_attributes-directory-2" > name="directory" > value="/mnt/dir2"/> \ > </instance_attributes> \ > </primitive> > ms msDRBD resDRBD \ > meta notify="true" > location locDRBD msDRBD \ > rule $id="locDRBD-rule" $role="Master" -inf: not_defined DRBD_Master > colocation col_FS_DRBD inf: resFS:Started msDRBD:Master > order ord_DRBD_FS inf: msDRBD:promote resFS:start > > corosync starts and both DRBDs are secondaries. The cluster also checks resFS > and answers: > > Failed actions > resFS_monitor_ (node=node1, call=13, rc=2, status=complete): invalid parameter
There's no interval for this action. Better file a bug and include a hb_report. > > The same happens reproducably when I cleanup the resource resFS. I debugged a > little bit a found out that pacemaker does not pass any parameters to the > resource agent. All variables OCF_RESKEY... are empty. > > When I add the attribute to the node (attrd_updater ...) are cleanup the > rescource resFS everything works as expected. > > I have pacemaker version 1.0.9-1 > > Is my problem the same that was described in bug #2458? No, thats for stop actions only. > Or is it something > new? _______________________________________________ Pacemaker mailing list: Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
