Troubleshooting Check List
- Make sure that you change both the jpa.databasePlatform and application.databaseType properties in the Cycle.properties file to match your database. Often one or both of these can be overlooked, causing exceptions to be thrown.
- Make sure to change the datasource.type property in the Cycle.properties file to jndi when running in a container.
- Make sure Cycle.properties is on the classpath and loaded by the container. You may see a message like "cannot find bundle Cycle" if it cannot find the properties file.
- Make sure that the coherence configuration files are on the classpath. Make sure that the well-known-address list in the cycle-coherence-config.xml and the OIPA coherence-config.xml match exactly if the OIPA instance is also used for cycle processing.
- Make sure that the cycle-coherence-cache-config.xml file is not shared by Cycle Agents. Each Cycle Agent should have its own cycle-coherence-cache-config.xml
- Make sure that the cycle-coherence-cache-config.xml file has unique identifiers as mentioned in the Coherence configuration section of this document.
Common Questions Concerning Cycle
The Cycle Agent doesn't do anything, even though I ran the Cycle Client.
The most common problem is that the Cycle Agent is not sharing the same Coherence Cluster as the other members of the Cycle group, or the Cycle Client. If the Cycle Agent is not sharing the same cluster, then it will not be able to receiving messages, and will not do anything. There are a few reasons why the Cycle Agent is not part of the same Coherence Cluster:
- Check the startup parameters for the application server to ensure that the system property is passed in. The system property is -Dtangosol.coherence.override=, and should point to the cycle-coherence-config.xml file. If the instance cannot find the cycle-coherence-config.xml file, it may start its own, and will be outside of the shared cluster.
- Check the well-known-addresses section in the cycle-coherence-config.xml file and make sure that the Cycle Agent and/or Cycle Client have exactly the same addresses listed. If you are denied access to the cluster, the instance may start its own, and will therefore be outside of the shared cluster.
- Make sure the cycle-coherence-cache-config.xml file is loaded. You can check this in the log file or the console output. Sometimes it will indicate that it could not be loaded, and load the default. If the default cache config file is loaded, it will not be in the shared coherence cluster.