Triggers and Active Cluster
Active Cluster handles triggers automatically. The timekeeper controls triggers, ensuring that a trigger occurs only as configured.
By default, when a trigger is executed, it is executed on each node in the cluster. However, for timer event triggers, the default is to execute the trigger once per cluster.
Active Cluster handles triggers in this way:
1. When a trigger fires in a TDV instance, the trigger manager checks to see if the trigger should be executed locally. Certain triggers, such as cache refreshes for the file cache, are always executed locally.
2. If a trigger is not local, the TDV instance that invokes the trigger is determined by the timekeeper based on a weighted round-robin policy. Each TDV instance in the cluster participates with a fixed weight of one (1) except for the timekeeper instance, whose weight is determined by the Cluster trigger distribution Weight of time keeper configuration setting. (The default is 1.)
You can set a timer event trigger to be invoked on every node in the cluster. See the TDV User Guide for information about how to set up triggers. Clear the “Only once per cluster” option to change the default trigger behavior.
You can also use these cluster system events as trigger conditions:
• ClusterServerConnected
• ClusterServerDisconnected
• ClusterServerShunned
See
Configuring an Active Cluster, for more information about configuring Active Cluster.