Cluster events

Edit online

This section describes how the PowerHA® SystemMirror® software responds to changes in a cluster to maintain high availability.

The PowerHA SystemMirror cluster software monitors all the components that make up the highly available application including disks, network interfaces, nodes and the applications themselves. The Cluster Manager uses different methods for monitoring different resources:

  • RSCT subsystem is responsible for monitoring networks and nodes.
  • The AIX® LVM subsystem produces error notifications for volume group quorum loss.
  • The Cluster Manager itself dispatches application monitors.

A PowerHA SystemMirror cluster environment is event-driven. An event is a change of status within a cluster that the Cluster Manager recognizes and processes. A cluster event can be triggered by a change affecting a network interface card, network, or node, or by the cluster reconfiguration process exceeding its time limit. When the Cluster Manager detects a change in cluster status, it executes a script designated to handle the event and its subevents.

Note: The logic of cluster events is described here for completeness. No manual intervention is required to ensure that PowerHA SystemMirror carries out cluster events correctly.

The following examples show some events the Cluster Manager recognizes:

  • node_up and node_up_complete events (a node is joining the cluster)
  • node_down and node_down_complete events (a node is leaving the cluster)
  • local or global network_down event (a network has failed)
  • global network_unstable event (a network is changing its state continuously)
  • global network_stable event (a network is no longer changing its state continuously)
  • network_up event (a network is connected)
  • swap_adapter event (a network adapter has failed and is swapped with a new adapter)
  • dynamic reconfiguration events
  • site_up and site_up_complete events (a site is joining the cluster)
  • site_down and site_down_complete events (a site is leaving the cluster)

When a cluster event occurs, the Cluster Manager runs the corresponding event script for that event. As the event script is being processed, a series of subevent scripts might be executed. The PowerHA SystemMirror software provides a script for each event and subevent. The default scripts are located in the /usr/es/sbin/cluster/events directory.

By default, the Cluster Manager calls the corresponding event script supplied with the PowerHA SystemMirror software for a specific event.

You can specify additional processing information to customize event handling for your site if needed.