How to implement a fence delay when using fence_vmware_soap as the fence agent in RHEL 5 or 6
Environment
- Red Hat Enterprise Linux (RHEL) 5 or 6 with the High Availability Add On
fence_vmware_soapis the configured fence agent for one or more nodes in/etc/cluster/cluster.conf
Issue
- How to add 'delay' option of fencing when fence_vmware_soap fence agent is being used ?
- In order to avoid fence race, 'delay' option of fencing is recommended in article Delaying Fencing in a Two Node Cluster to Prevent Fence Races or "Fence Death" Scenarios. How to use for fence_vmware_soap fence agent ?
- I have configured a
delayforfence_vmware_soap, but its not taking effect. - Two different clusters with different releases of
fence-agentsbehave differently with respect to thedelayspecified forfence_vmware_soap. One delays, one doesn't.
Resolution
RHEL 6
Update to [`fence-agents-3.1.5-35.el6_5.3`](http://rhn.redhat.com/errata/RHBA-2014-0141.html) or later, and set the `delay` attribute for `fence_vmware_soap` to the desired value.
RHEL 5
Update to [`cman-2.0.115-118.el5_10.4`](http://rhn.redhat.com/errata/RHBA-2014-0282.html) or later, and set the `delay` attribute for `fence_vmware_soap` to the desired value.
- Below is an example configuration which shows how to add delay in vmware fencing.
<clusternode name="node1.example.com" nodeid="1" votes="1">
<fence>
<method name="1">
<device name="Vmware" port="node1" uuid="4223dbb3-6ec6-fg65-ef4c-dhu7562dff56cf" delay="10"/>
</method>
</fence>
</clusternode>
- After the above modifications increase the cluster config version by 1 and propagate the changes to all the nodes.
For pacemaker clusters then see the following solution: How do I delay fencing to prevent fence races when using a shared stonith device?
Root Cause
The fence_vmware_soap fence agent did not respect the delay attribute in certain releases, which could cause a situation where two nodes fenced each other at the same time. With the above listed updates, the agent waits for the given amount of time as expected, and race conditions no longer occur.
Red Hat modified fence_vmware_soap via Bugzilla #1051159 on RHEL 6 to allow for the specification of a delay and Bugzilla #1073821 for RHEL 5.
This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.