SUSE Linux Enterprise High Availability Extension 11
SUSE Linux Enterprise Server 11
Filesystem Clone Resources

A clone resource in the cluster has filesystem primitives defined. However, when any one node is rebooted, all the filesystems in the cloned resource are unmounted and then mounted again when the rebooted node joins the cluster. In a three node cluster, the /clonefs filesystem in the example below would be unmounted on all nodes when any one node was rebooted, and then be mounted again on all nodes when the rebooted node joined the cluster.

The filesystem clone resource configuration looks like:

<clone id="filesystem-clone-001">
  <meta_attributes id="filesystem-clone-001-meta_attributes">
    <nvpair id="filesystem-clone-001-meta_attributes-target-role" name="target-role" value="Started"/>
  </meta_attributes>
  <primitive class="ocf" id="ocfs2-filesystem" provider="heartbeat" type="Filesystem">
    <operations id="ocfs2-filesystem-operations">
      <op id="ocfs2-filesystem-op-monitor-20" interval="20" name="monitor" timeout="40"/>
    </operations>
    <instance_attributes id="ocfs2-filesystem-instance_attributes">
      <nvpair id="ocfs2-filesystem-instance_attributes-device" name="device" value="/dev/disk/by-path/iqn.2012-11.lab:north1g-lun-5-part2"/>
      <nvpair id="ocfs2-filesystem-instance_attributes-directory" name="directory" value="/clonefs"/>
      <nvpair id="ocfs2-filesystem-instance_attributes-fstype" name="fstype" value="ocfs2"/>
    </instance_attributes>
  </primitive>
</clone>

or configured like:

<clone id="filesystem-clone-001">
  <meta_attributes id="filesystem-clone-001-meta_attributes">
    <nvpair id="filesystem-clone-001-meta_attributes-target-role" name="target-role" value="Started"/>
    <nvpair id="filesystem-clone-001-meta_attributes-interleave" name="interleave" value="false"/>
  </meta_attributes>
<snip/>

Set the cluster clone metadata interleave to true. Referring to the example above, run:

crm_resource --resource filesystem-clone-001 --set-parameter interleave --meta --parameter-value true

Once you have set interleave to true, use cibadmin -Q to review the changes. Your clone resource metadata should look like:

<clone id="filesystem-clone-001">
  <meta_attributes id="filesystem-clone-001-meta_attributes">
    <nvpair id="filesystem-clone-001-meta_attributes-target-role" name="target-role" value="Started"/>
    <nvpair id="filesystem-clone-001-meta_attributes-interleave" name="interleave" value="true"/>
  </meta_attributes>
  <primitive class="ocf" id="ocfs2-filesystem" provider="heartbeat" type="Filesystem">
    <operations id="ocfs2-filesystem-operations">
      <op id="ocfs2-filesystem-op-monitor-20" interval="20" name="monitor" timeout="40"/>
    </operations>
    <instance_attributes id="ocfs2-filesystem-instance_attributes">
      <nvpair id="ocfs2-filesystem-instance_attributes-device" name="device" value="/dev/disk/by-path/iqn.2012-11.lab:north1g-lun-5-part2"/>
      <nvpair id="ocfs2-filesystem-instance_attributes-directory" name="directory" value="/clonefs"/>
      <nvpair id="ocfs2-filesystem-instance_attributes-fstype" name="fstype" value="ocfs2"/>
    </instance_attributes>
  </primitive>
</clone>

The default interleave value is false. When the interleave is set to false, all nodes are expected to unmount and then remount the clone filesytems when a node is rebooted. If other types of resources are included in the clone group with the filesystem resources, the interleave does not affect the group. It mostly just affects filesystem only clone or clone group resources.