IJ35784: CRONTASK WITH A FREQUENCY SET TO "EVERY 3 MONTH(S)" OR EVEN MORE WILL NOT RUN

APAR status

• Closed as program error.

Error description

• ERROR DESCRIPTION:
  Crontask with a frequency set to "Every 3 month(s)" or even
  more will not run
  
  PROBLEM:
  You may face the scenario where you no longer receiving some of
  the scheduled reports (specifically the ones that are running
  on a 3 or 4 month increments.)
  
  Using a frequency for the crontask as 1 week, or 1 month, makes
  the REPORTSCHEDULE crontask run fine.
  
  However when the REPORTSCHEDULE crontask name has a frequency
  of 3 months or more, it simply does not run, no errors in the
  log.
  
  This problem can be simulated by setting a cron task instance
  to run every 10 minutes, reproducible steps:
  
  1. Take an existing crontask, for example RemoveDeadSessions.
  Remove all its active instances and cron task history. You can
  do it in the back end:
  delete from taskscheduler where taskname like
  'RemoveDeadSessions%';
  delete from  crontaskhistory where crontaskname like
  'RemoveDeadSessions%';
  COMMIT;
  
  2. Create a new instance, set the schedule to be every 10
  minutes.
  
  3. Before the 10 minutes is up, stop the server
  
  4. Restart server, wait a few minutes and stop the server.
  Repeat this step as many times as you want.  You will see that
  the cron task instance is never run.  This is because every
  time a server restart happens, the cron task is scheduled to
  run 10 minutes later but the server gets restarted before it
  reaches 10 minutes so the cron task never gets to run.  In
  reality, this only happens to cron tasks that are scheduled to
  run months or years apart as the chances of the server being
  restarted within that timeframe is high.
  
  This problem can be worked around by manually setting the
  taskscheduler.lastrun field after such a cron task instance is
  created, or have an automation script to do that.  Then it will
  be scheduled as if it was last run at the time specified in
  that field. To fix this in the code, if cron task has never
  run, will use its reload request time
  (crontaskinstance.reloadreqtime, instead of the null
  taskscheduler.lastrun) as the starting time to calculate the
  next run schedule.
  
  STEPS TO REPLICATE :
  1. Open maximo as maxadmin user
  2. Ensure you have a crontask (REPORTSCHEDULE) with the
  frequency set to "Every 3 month(s)" or even more
  3. The crontask will never run
  
  OR
  
  This problem can be simulated by setting a cron task instance
  to run every 10 minutes, reproducible steps:
  
  1. Take an existing crontask, for example RemoveDeadSessions.
  Remove all its active instances and cron task history. You can
  do it in the back end:
  delete from taskscheduler where taskname like
  'RemoveDeadSessions%';
  delete from  crontaskhistory where crontaskname like
  'RemoveDeadSessions%';
  COMMIT;
  
  2. Create a new instance, set the schedule to be every 10
  minutes.
  
  3. Before the 10 minutes is up, stop the server
  
  4. Restart server, wait a few minutes and stop the server.
  Repeat this step as many times as you want.  You will see that
  the cron task instance is never run.  This is because every
  time a server restart happens, the cron task is scheduled to
  run 10 minutes later but the server gets restarted before it
  reaches 10 minutes so the cron task never gets to run.  In
  reality, this only happens to cron tasks that are scheduled to
  run months or years apart as the chances of the server being
  restarted within that timeframe is high.
  
  This problem can be worked around by manually setting the
  taskscheduler.lastrun field after such a cron task instance is
  created, or have an automation script to do that.  Then it will
  be scheduled as if it was last run at the time specified in
  that field. To fix this in the code, if cron task has never
  run, will use its reload request time
  (crontaskinstance.reloadreqtime, instead of the null
  taskscheduler.lastrun) as the starting time to calculate the
  next run schedule.
  
  NOTE : If the same crontask is to run every 1 week, or every 1
  month, it works OK.
  
  
  CURRENT ERRONEOUS RESULT:
  The crontask will never run
  
  EXPECTED RESULT:
  The crontask will run
  
  ENVIRONMENT (SYSTEM INFO):
  App Server IBM WebSphere Application Server 9.0.0.10
  Tivoli's process automation engine 7.6.1.2-IFIX20210104-1913
  Build 20200715-0100 DB Build V7612-284 HFDB Build HF7612-15
  IBM Maximo Asset Management Work Centers 7.6.0.4 Build
  20200715-0100 DB Build V7604-119 HFDB Build HF7680-10
  IBM Maximo Asset Management Scheduler 7.6.8.0 Build
  20200715-0100 DB Build V7680-148 HFDB Build HF7680-15
  IBM Tpae Integration Framework
  7.6.1.2-MIF_7612_IFIX.20200909-1113 Build 20200715-1444 DB
  Build V7612-11
  IBM Maximo Asset Management 7.6.1.2 Build 20200715-0100 DB
  Build V7611-01
  IBM Maximo for Oil and Gas 7.6.1.0-20200805-1125 Build
  20161020-0803 DB Build V7610-139 HFDB Build HF7610-56
  IoT Connection Utility 7.6.0.3 Build 20200701-1409 DB Build
  V7603-32
  Ontracks MaxEdit 7.6.0.1 Build 20180910-1005 DB Build V7601-01
  Server OS Linux 4.4.0-1114-aws
  Server DB DB2/LINUXX8664 11.1 (SQL11014)
  

Local fix

• This problem can be simulated by setting a cron task instance
  to run every 10 minutes, reproducible steps:
  
  
  
  1. Take an existing crontask, for example RemoveDeadSessions.
  Remove all its active instances and cron task history. You can
  do it in the back end:
  
  
  
  delete from taskscheduler where taskname like
  'RemoveDeadSessions%';
  
  delete from  crontaskhistory where crontaskname like
  'RemoveDeadSessions%';
  
  COMMIT;
  
  
  
  2. Create a new instance, set the schedule to be every 10
  minutes.
  
  
  
  3. Before the 10 minutes is up, stop the server
  
  
  
  4. Restart server, wait a few minutes and stop the server.
  Repeat this step as many times as you want.  You will see that
  the cron task instance is never run.  This is because every
  time a server restart happens, the cron task is scheduled to
  run 10 minutes later but the server gets restarted before it
  reaches 10 minutes so the cron task never gets to run.  In
  reality, this only happens to cron tasks that are scheduled to
  run months or years apart as the chances of the server being
  restarted within that timeframe is high.
  
  
  
  
  
  This problem can be worked around by manually setting the
  taskscheduler.lastrun field after such a cron task instance is
  created, or have an automation script to do that.  Then it will
  be scheduled as if it was last run at the time specified in
  that field.
  
  
  
  To fix this in the code, if cron task has never run, will use
  its reload request time (crontaskinstance.reloadreqtime,
  instead of the null taskscheduler.lastrun) as the starting time
  to calculate the next run schedule.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * MAXIMO                                                       *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * When Application is restarted every month, crontasks which   *
  * are scheduled to run every 2 or more months never run.       *
  ****************************************************************
  

Problem conclusion

• The fix for this APAR will be contained in a future release.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  SYSTEM CONFIG

• Fixed component ID

  5724R46S1

Applicable component levels