Detecting loops between jobs and job streams in current and trial plans

Ensure robust detection of loops involving nodes (that is, jobs and job streams) in both current and trial plans.

To detect loops in jobs and job streams, perform the following steps:
  1. Stop WebSphere Application Server Liberty Base as described in Application server - starting and stopping.
  2. Browse to the TWSConfig.properties file, located in the following paths, depending on your operating system:
    On Windows operating systems
    <TWA_home>\usr\servers\engineServer\resources\properties
    On UNIX operating systems
    <TWA_DATA_DIR>/usr/servers/engineServer/resources/properties
  3. Set the com.ibm.tws.planner.monitor.verboseLoopLogging property to true.
  4. Restart WebSphere Application Server Liberty Base as described in Application server - starting and stopping.
  5. Monitor the trace.log file when generating the current plan, adding dependencies, generating a forecast plan, or performing other actions which might generate a loop. The file is located in the following paths, depending on your operating system:
    On Windows operating systems
    <TWA_home>\TWS\stdlist\appserver\engineServer\logs
    On UNIX operating systems
    <TWA_DATA_DIR>/stdlist/appserver/engineServer/logs
  6. Messages similar to the following are displayed:
    In plan "your_plan", the node "your_node" is inside a loop
    In plan "your_plan", there are n plans inside a loop

To reduce verbosity, loop notifications are only logged when a change occurs—such as a job or job stream entering or exiting a detected loop.