DB2 10.5 for Linux, UNIX, and Windows

Detecting an unplanned outage

Before you can respond to the failure of a component, you must detect that the component failed. DB2® Data Server has several tools for monitoring the health of a database, or otherwise detecting that a database has failed. You can configure these tools to notify you or take predefined actions when they detect a failure.

Procedure

You can use the following tools to detect when a failure has occurred in some part of your DB2 database solution:

DB2 fault monitor facility: The DB2 fault monitor facility keeps DB2 database instances up and running. When the DB2 database instance to which a DB2 fault monitor is attached terminates unexpectedly, the DB2 fault monitor restarts the instance. If your database solution is implemented in a cluster, you should configure the cluster managing software to restart failed database instances instead of the DB2 fault monitor.
Heartbeat monitoring in clustered environments: Cluster managing software uses heartbeat messages between the nodes of a cluster to monitor the health of the nodes. The cluster manager detects that a node has failed when the node stops responding or sending any messages.
Monitoring DB2 High Availability Disaster Recovery (HADR) databases: The HADR feature has its own heartbeat monitor. The primary database and the standby database each expect heartbeat messages from the other at regular intervals.