Configuring Analytic Engine High Availability
You can configure the Analytic JVM for automatic restart if the JVM stops unexpectedly, or experiences certain errors. You configure these settings by adding parameters to the wrapper.conf file in the same way you configure memory allocation, as described in
Configuring the Java Service Wrapper.
The Java Wrapper Service restarts the Analytic Engine in the following instances:
The JVM terminates unexpectedly. This occurs automatically and you do not need to specify any parameters for this behavior.
A thread dead lock is detected. Specify the deadlock interval parameter to define the interval at which dead lock detection is performed:
wrapper.check.deadlock.interval=xxx where
xxx is the interval length in seconds. The default value is 60 seconds.
A Heap or Perm out of memory error occurs. This is set by default and you do not need to specify any parameters for this behavior. It requires the additional parameter
-XX:+HeapDumpOnOutOfMemoryError set by default as a
wrapper.java.additional.<n>=<parameter> parameter.
Note:
In case of an out of memory error, the JVM creates a heap dump in the Software AG_directory /optimize/analysis/bin directory (java_pidXXXX.hprof).
The JVM does not respond - In normal operation, the Wrapper pings the JVM once every 90 seconds to make sure that its process has not frozen up. The ping timeout (300 seconds by default) is the amount of time to allow for a response from the JVM before the Wrapper assumes that it is not responding and restarts it. These values are specified by the
wrapper.ping.interval and the
wrapper.ping.timeout parameters.
The JVM is terminated if the JVM does not shut down as expected. If the JVM does not shut down during the interval specified in the wrapper.shutdown.timeout parameter and the wrapper.jvm_exit.timeout, the JVM is terminated.