[Pacemaker] Resource migrates between nodes with equal scores for resource

Jesse W. Hathaway jesse at mbuki-mvuki.org
Thu Aug 14 14:23:32 EDT 2008

I have two nodes hatest01 & hatest02 and a resource group 
ESB_HA with equal scores on both nodes:


  1. Both nodes are active and ESB_HA is running on hatest01

  2. I shutdown down heartbeat on hatest01, ESB_HA migrates 
     to hatest02

  3. I restart heartbeat on hatest01, ESB_HA migrates to 

Given that both nodes have equal scores why does step 3

thanks for the help, Jesse


Resource            Score     Node            Stickiness #Fail
ESB_APACHE          300       hatest01        0          -50                      
ESB_APACHE          300       hatest03        0          -50                      
ESB_HA              0         hatest01        0          -50                      
ESB_HA              0         hatest03        0          -50                      
ESB_IP              300       hatest01        0          -50                      
ESB_IP              300       hatest03        0          -50                      
ESB_SYNAPSE         300       hatest01        0          -50                      
ESB_SYNAPSE         300       hatest03        0          -50                      

<?xml version="1.0"?>
<cib admin_epoch="2008081401" epoch="0" num_updates="0" generated="false" have_quorum="true">
      <cluster_property_set id="cib-bootstrap-options">
          <nvpair id="cp-1" name="no-quorum-policy" value="ignore"/>
          <nvpair id="cp-3" name="default-resource-failure-stickiness" value="-50"/>
      <node id="2860975f-746b-4ad6-92d1-5c07f9ec1717" uname="hatest01" type="normal"/>
      <node id="9b4a3e1e-7537-492f-bea8-d6c08f328216" uname="hatest03" type="normal"/>
      <group id="ESB_HA">
        <primitive class="ocf" id="ESB_IP" type="IPaddr" provider="heartbeat">
            <op id="ESB_IP_mon" interval="15s" name="monitor" timeout="10s"/>
          <instance_attributes id="ESB_IP_attr">
              <nvpair id="ESB_IP_ip01" name="ip" value=""/>
              <nvpair id="ESB_IP_netmask01" name="cidr_netmask" value="24"/>
              <nvpair id="ESB_IP_nic01" name="nic" value="eth0"/>
        <primitive class="lsb" id="ESB_APACHE" type="apache2-ha" provider="heartbeat">
            <op id="op-1" name="start" timeout="60s" on_fail="restart"/>
            <op id="op-2" name="monitor" interval="10s" timeout="5s" on_fail="restart"/>
            <op id="op-3" name="stop" timeout="20s" on_fail="restart"/>
        <primitive class="lsb" id="ESB_SYNAPSE" type="wso2-esb-ha" provider="heartbeat">
            <op id="op-4" name="start" timeout="60s" on_fail="restart"/>
            <op id="op-5" name="monitor" interval="10s" timeout="10s" on_fail="restart"/>
            <op id="op-6" name="stop" timeout="20s" on_fail="restart"/>
      <rsc_location id="loc-1" rsc="ESB_IP" node="hatest01" score="300"/>
      <rsc_location id="loc-2" rsc="ESB_IP" node="hatest03" score="300"/>
      <rsc_location id="loc-3" rsc="ESB_SYNAPSE" node="hatest01" score="300"/>
      <rsc_location id="loc-4" rsc="ESB_SYNAPSE" node="hatest03" score="300"/>
      <rsc_location id="loc-5" rsc="ESB_APACHE" node="hatest01" score="300"/>
      <rsc_location id="loc-6" rsc="ESB_APACHE" node="hatest03" score="300"/>

More information about the Pacemaker mailing list