Is there a way to have a RHEL 6 node in a two-node cluster panic or reboot itself when there's a problem instead of needing power or storage based fencing?

Solution Unverified - Updated

Environment

  • Red Hat Enterprise Linux (RHEL) 6 with the High Availability Add On
  • cman or cman/pacemaker cluster
  • Two-node cluster
  • No direct access to shared SCSI storage or power fencing mechanisms

Issue

  • Can I configure cluster nodes to use panics or reboots of some kind as a fence mechanism?
  • I don't have access to power fencing or shared storage. How can I configure the cluster nodes to fence themselves?

Resolution

Currently there is no fence mechanism for two-node clusters that do not have access to shared SCSI storage or power control. If this is a use case that is applicable to your environment, please contact Red Hat Global Support Services for additional information on potential solutions.

Root Cause

Typically when power-fencing is unavailable, storage-based fencing like fence_scsi, fence_brocade, or other fiber-channel-switch mechanisms would be used. However, this requires the use of shared SCSI storage typically from a storage array, so environments without such storage (such as if the nodes use NFS or other non-SCSI network-based storage) may not have any available fence mechanism to use.

There is an alternative method called fence_sanlock which is a Technology Preview in RHEL 6 Update 5, but it requires shared storage in the two-node case and requires a hardware watchdog for all use cases.

SBR
Product(s)
Components

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.