Failover to the standby management node

In an IBM® Spectrum Cluster Foundation Community Edition high availability environment, failover occurs when the active management node is no longer available, and high availability services and virtual IP addresses are automatically migrated to the standby management node. The standby management node becomes the active management node after all high availability services are started.

Failover has two modes: manual and automatic. By default, the failover mode is set to automatic. In automatic failover mode, a failover is automatically triggered if any of the following conditions are met:
  • The management node is accidentally powered off.
  • The management node is rebooted and cannot boot in time.
  • The management node hardware fails, and machine is not operational.
  • The management node connection to the provision network is down.
  • The management node high availability manager is down.
It can take a couple minutes for a condition to be detected by the high availability manager.

In manual failover mode, the failover does not occur automatically. Manual mode can be used for cluster maintenance or to force a failover. For example, if you stop IBM Spectrum Cluster Foundation Community Edition services in a high availability environment, the services automatically move to the failover node. For maintenance operations, first set the failover mode to manual and then stop IBM Spectrum Cluster Foundation Community Edition services. Stop and start IBM Spectrum Cluster Foundation Community Edition services using the service pcm start command and service pcm stop command.

To set the failover mode to automatic, specify the following command: pcmhatool failmode -m auto. Before you change the failover mode to automatic, ensure that no errors exist on the standby node, by issuing the pcmhatool check command.

Note: The virtual IP address is not available when IBM Spectrum Cluster Foundation Community Edition services are stopped. It can take a few minutes for the virtual IP address to become available and services to resume after a failover occurs.