So we are using a failover cluster to make some running java-applications highly available. And we are using java service wrapper to run the java-processes in out windows environment). We are having an issue that, although the cluster says that the java-application has failed, the Java-application is still running.
This happens when for example the resource disk that java-applications use fails and the cluster resources who are dependant of this also fail automaticly. The cluster is responsible for this proces.
The thing is that the cluster doesn't seem to be communicating properly with the java-applications.
i want to thorougly document and research this issue, as it will lead to cause a failover, which will lead to double running processes on the cluster. (with dataloss or datacorruption as a consequence).
Now I want something so I can monitor the status of the JVM's.
using task manager I already know(just checking for Java.exe), but i know that it should be also possible to monitor the JVM's using monitorsoftwrae.
Thing that i need to mention is that there isn't a java version installed on the servernodes. We use the wrapper software to point to some JRE sources to run the application. A JDK is also not installed.
If i really need to i will install a JDK or JRE on it for testingpurposes.