• Implement the shutdown process as described at

http://wiki.trick.ca/linux-ha/ClusterResourceManagerDaemon

• Added disconnect actions for LRM, HA and CCM
• Timer helpers report success or failure
• Restart timers only if they were running to begin with
• Nodes now request the DC to shut them down
• The new process allows the escalation timer to work (since control is returned to GMainLoop code)
• Added a hook for adding/removing actions on a change of state
• Dont announce ourselves if we're not ready, healthy or shutting down
• Dont vote if we've asked to shut down

CVS patchset: 2820
CVS date: 2004/04/12 15:34:50

Mercurial revision: 880b7e6d34853073dc66e2a3613238a2f5d12a2e