Just FYI, this configuration also causes shutdown problems for pacemaker.  When 
I stop openais while running this configuration, lrmd and pengine are orphaned 
and become children of init (process 1).  Maybe this is why the resources never 
stop.

By the way, I noticed that I had the score designator it the wrong place in the 
collocation area, and it should be:

<rsc_colocation id="coloc-1" score="INFINITY">
        <resource_set id="coloc-1-set-1" sequential="true">
                <resource_ref id="world1" /> 
                <resource_ref id="world2" /> 
        </resource_set>
</rsc_colocation>

Unfortunately, that didn't change the behavior.

-Frank


-Frank


> -----Original Message-----
> From: Frank DiMeo [mailto:[email protected]]
> Sent: Tuesday, November 24, 2009 6:00 PM
> To: [email protected]
> Subject: [Pacemaker] another basic configuration question
> 
> Hi All,
> 
> I'm using the enclosed .xml file to control the startup sequence and
> collocation of two resources.  When I use this xml file, I get the
> following behavior:
> 
> 1) start both nodes, resources world1 and world2 start up in the
> correct sequence on ubuntu_2
> 2) execute crm_standby -U ubuntu_2 v -on, world2 and world2 stay on
> ubuntu_2.  This seems odd.
> 3) execute crm_standby -U ubuntu_2 v -off, world2 and world2 stay on
> ubuntu_2.
> 4) execute crm_resource --resource world1 --move -h ubuntu_1, world1
> and world2 stay on ubuntu_2.  Again, I thought world1 could move, and
> if it did, would take world2 with it.
> 5) execute crm_resource --resource world1 --un-move -h ubuntu_1
> 6) try to move world2, it won't move either, but this is not unexpected
> given the collocation constraint.
> 7) stop pacemaker (via /etc/init.d/openais stop) on ubuntu_2, and after
> about a minute, corosync et al stop, but world1 and world2 still
> continue to run on ubuntu_2, and both resources ALSO start running on
> ubuntu_1.
> 
> I've checked my resources using ocf-tester on both nodes (both pass on
> both nodes), and also manually verified both RA's on both nodes.  Does
> this behavior make sense given the xml file?
> 
> I've also run both of these resources using ONLY the collocation
> constraint (without the ordering constraint), the they seem to behave
> as I'd expect (i.e. they run on the same node together even when
> forcibly moved, or when a node is put into standby), as long as I
> specify "sequential=true" in the collocation constraint.  However, if I
> specify "sequential=false" in the collocation constraint, each resource
> tries to execute on the node opposite of the one that's running the
> other resource.
> 
> Thanks again in advance,
> Frank
<cib validate-with="pacemaker-1.0" crm_feature_set="3.0.1" have-quorum="0" admin_epoch="1" epoch="8" num_updates="1" cib-last-written="Fri Nov 20 07:46:43 2009" dc-uuid="ubuntu_2">
  <configuration>
    <crm_config>
      <cluster_property_set id="cib-bootstrap-options">
        <nvpair id="option-1" name="symmetric-cluster" value="true"/>
        <nvpair id="option-2" name="no-quorum-policy" value="ignore"/>
        <nvpair id="option-3" name="stonith-enabled" value="false"/>
        <nvpair id="option-4" name="cluster-delay" value="5s"/>
        <nvpair id="cib-bootstrap-options-cluster-infrastructure" name="cluster-infrastructure" value="openais"/>
        <nvpair id="cib-bootstrap-options-expected-quorum-votes" name="expected-quorum-votes" value="2"/>
        <nvpair id="cib-bootstrap-options-dc-deadtime" name="dc-deadtime" value="10s"/>
        <nvpair id="cib-bootstrap-options-election-timeout" name="election-timeout" value="10s"/>
      </cluster_property_set>
    </crm_config>
    <nodes>
      <node id="ubuntu_2" uname="ubuntu_2" type="normal"/>
      <node id="ubuntu_1" uname="ubuntu_1" type="normal"/>
    </nodes>
    <resources>
      <primitive id="world1" class="ocf" type="world1test" provider="bbnd">
        <operations>
          <op id="world1check" name="monitor" interval="10s" requires="nothing" on-fail="restart"/>
        </operations>
      </primitive>
      <primitive id="world2" class="ocf" type="world2test" provider="bbnd">
        <operations>
          <op id="world2check" name="monitor" interval="10s" requires="nothing" on-fail="restart"/>
        </operations>
      </primitive>
    </resources>
    <constraints>
      <rsc_order id="order-1">
	<resource_set id="order-1-set-1" sequential="true">
	  <resource_ref id="world1"/>
	  <resource_ref id="world2"/>
	</resource_set>
      </rsc_order>
      <rsc_colocation id="coloc-1">
	<resource_set id="coloc-1-set-1" sequential="true" score="INFINITY">
	  <resource_ref id="world1"/>
	  <resource_ref id="world2"/>
	</resource_set>
      </rsc_colocation>
    </constraints>
  </configuration>
</cib>
_______________________________________________
Pacemaker mailing list
[email protected]
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
_______________________________________________
Pacemaker mailing list
[email protected]
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Reply via email to