- During recovery, exceptions are caught when trying to recover a workflow so that incorrect configuration of one workflow cannot lead to the entire recovery process failing.
- When stopping the application in a normal way, the engines will first finish all current tasks before shutting down to avoid tasks being interrupted midway. (See Clean Engine shutdown)
- The name of the engineNode will be based on the
getName()method of the
java.lang.management.RuntimeMXBeanclass, which in most cases combines the hostname and process id to ensure the name is unique across different nodes on the same hostname.
- When one of the nodes is unresponsive for a long time (read; did not update it’s lastActive field for a time greater than the defined timeout) an error message will be shown:
ERROR: node='<engineNode.name>' appears inactive, perhaps it has shut down unexpectedly? lastActive='<engineNode.lastActive>', previousStatus='<engineNode.status>', status='NOT_RESPONDING'
- Prevented a nullpointerException if one of the nodes is cleaned and afterwards regains connection to the database.
Look at the
net.democritus.workflow.EngineNodeConfigclass to see which parameters can be modified.
- baseComponents: 22.214.171.124 -> 126.96.36.199
- expanders: 188.8.131.52 -> 184.108.40.206