Understanding Oracle Solaris Cluster HA for Apache Tomcat Fault Monitor

This section describes the Oracle Solaris Cluster HA for Apache Tomcat fault monitor's probing algorithm and functionality, as well as the conditions, messages, and recovery actions associated with unsuccessful probing.

For conceptual information on fault monitors, see Oracle Solaris Cluster Concepts Guide.

Resource Properties

The Oracle Solaris Cluster HA for Apache Tomcat fault monitor uses the same resource properties as the resource type SUNW.gds. Refer to the SUNW.gds(5) man page for a complete list of resource properties.

Probing Algorithm and Functionality

The following steps are executed to monitor the sanity of Apache Tomcat.

1. Sleeps for Thorough_probe_interval.

2. Pings the Host, which is configured in the Apache Tomcat HA parameter file unless the TestUrl parameter is specified in the Tomcat parameter file.

3. Connects to the Apache Tomcat via Host and Port. If the connection is successful, it sends the TestCmd and tests whether the ReturnString comes back. If it fails, it is rescheduled after five seconds. If this fails again, the probe restarts the Apache Tomcat.

   ---

   Caution - The ReturnString cannot be Connection refused because this string is returned if no connection is possible.

   ---

4. If the Apache Tomcat process has died, pmf will interrupt the probe to immediately restart the Apache Tomcat.

5. If the Apache Tomcat is repeatedly restarted and subsequently exhausts the Retry_count within the Retry_interval, then a failover is initiated for the resource group onto another node. This is done if the resource property Failover_enabled is set to TRUE.