JobScheduler crashes on setting order state to end node for API jobs

Current Situation

• The JobScheduler crashes when any API job is used with a Monitor script to set the state of an Order to an end node of the job chain by use of the spooler_task.order.state property.
• The error message reported with the scheduler.log file is as follows:

  .16 23:33:31.772 scheduler 83750 16632.19B0 9953.563MB TerminateProcess(0000000000000CA4,99)  
  .16 23:33:31.773 scheduler    0 16632.19B0 9953.563MB [xc.insert 1, "java.lang.IllegalStateException: Not in C++ thread. This is 'Thread[C++ Heart_beat_watchdog_thread,5,main]', expected is 'Thread[C++,5,main]'"]
  .16 23:33:31.773 scheduler    0 16632.19B0 9953.563MB [xc.insert 2, "CallVoidMethodA"]
  .16 23:33:31.773 scheduler    0 16632.19B0 9953.563MB {scheduler} start_process(""C:\Program Files\sos-berlin.com\jobscheduler\apmacwin_4444\bin\scheduler.exe" -service=sos_scheduler_apmacwin_4444 -exclusive -id=apmacwin_4444 -sos.ini=C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444/config/sos.ini -config=C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444/config/scheduler.xml -ini=C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444/config/factory.ini -env="SCHEDULER_HOME=C:/Program Files/sos-berlin.com/jobscheduler/apmacwin_4444" -env=SCHEDULER_DATA=C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444 -param=C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444 -cd=C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444 -include-path=C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444 -pid-file=C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444/logs/scheduler.pid -renew-service "-renew-spooler=C:\Program Files\sos-berlin.com\jobscheduler\apmacwin_4444\bin\scheduler.exe"")
  .16 23:33:31.776 scheduler    0 16632.19B0 9953.563MB close(4) C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444/logs/scheduler.pid
  .16 23:33:31.780 scheduler   15 16632.19B0 9953.563MB unlink("C:/ProgramData/sos-berlin.com/jobscheduler/apmacwin_4444/logs/scheduler.pid")
  .16 23:33:31.780 scheduler    0 16632.19B0 9953.563MB {scheduler} TerminateProcess( GetCurrentProcess() );
  

• The monitor script to reproduce the problem is as follows:

  function spooler_process_before() 
  {
    spooler_task.order.state = 'success';
    return false;
  }
  

• Observations
  • The problem occurs independently from the number of job nodes in a job chain.
  • The following jobs are reported to work and not to reproduce the problem:
    • Shell jobs using or not using the above Monitor script:

      <job order="yes" stop_on_error="false">
        <script language="shell"><![CDATA[echo "hello world"
      ]]></script>
        <monitor name="process0" ordering="0">
          <script language="java:javascript"><![CDATA[function spooler_process_before() 
      {
        spooler_task.order.state = 'success';
        return false;
      }
      ]]></script>
        </monitor>
        <run_time/>
      </job>

    • Java jobs not using the above Monitor script:

      <job order="yes" stop_on_error="false">
        <params>
          <param name="to" value="spam@sos-berlin.com"/>
          <param name="subject" value="JITL-606"/>
        </params>
        <script language="java" java_class="sos.scheduler.managed.JobSchedulerManagedMailJob"/>
        <run_time/>
      </job>

    • JavaScript jobs not using the above Monitor script:

      <job order="yes" stop_on_error="false">
        <script language="java:javascript"><![CDATA[function spooler_process_before() 
      {
        spooler_task.order.state = 'success';
        return false;
      }
      
      function spooler_process()
      {
        return true;
      }
      ]]></script>
        <run_time/>
      </job>

  • The following jobs using the above Monitor script reproduce the problem:
    • Java jobs using the above Monitor script:

      <job order="yes" stop_on_error="false">
        <params>
          <param name="to" value="spam@sos-berlin.com"/>
          <param name="subject" value="JITL-606"/>
        </params>
        <script language="java" java_class="sos.scheduler.managed.JobSchedulerManagedMailJob"/>
        <run_time/>
      </job>

    • JavaScript jobs using the above Monitor script:

      <job order="yes" stop_on_error="false">
        <script language="java:javascript"><![CDATA[function spooler_process()
      {
        return true;
      }
      ]]></script>
        <monitor name="process0" ordering="0">
          <script language="java:javascript"><![CDATA[function spooler_process_before() 
      {
        spooler_task.order.state = 'success';
        return false;
      }
      ]]></script>
        </monitor>
        <run_time/>
      </job>

How to reproduce

• Unzip the attached Job Configuration JITL-606.zip. The attached configuration uses JobSchedulerManagedMailJob to reproduce the issue.
• Login to the JOC Cockpit.
• Navigate to the Job Chains tab and start the order test_wizard.
• The JobScheduler crashes after some time.

Desired Behavior

• The JobScheduler should not crash even if the order state is set to an end node using a Monitor script for an API job in a job chain.