IV94649: SCHEDULED JOB RUNS FROM MULTIPLE JVMS WHEN IT IS ONLY SUPPOSED TO RUN ONE TIME.

APAR status

• Closed as program error.

Error description

• Problem Description:
  Scheduled Job runs from Multiple JVMs when it is only supposed
  to run one time.
  
  
  Steps to Reproduce:
  Run the same memory intensive escalation instance multiple,
  overlapping times.
  
  
  CURRENT ERRONEOUS RESULT:
  All of the jobs fail.
  
  
  EXPECTED RESULT:
  Single jobs runs to successful completion.
  
  
  ADDITIONAL INFORMATION
  The reason for this issue is the overlapping cron jobs. When
  such overlap happens, you would have multiple server entries for
  the same task names. The update query constructuted to update
  the taskschedule jobs looks like (Note that I have simplified
  the query to illustrate the issue)
  
  update taskscheduler set servertimestamp = TIMESTAMP ,
  servername = 'SERVER1' where servername = 'SERVER1' and taskname
  in ( 'TASK1', 'TASK2' ) or taskname in ('TASK3', 'TASK4' ) or
  taskname in ( 'TASK5', 'TASK6' );
  
  which when someone casually looks at it, can easily ignore and
  not see an issue. The problem is, becasue of the overlapping
  tasks running on multiple servers, the above query yields
  records for other servers and not just for SERVER1.
  
  The correct query should have been
  
  update taskscheduler set servertimestamp = TIMESTAMP ,
  servername = 'SERVER1' where servername = 'SERVER1' and
  (taskname in ( 'TASK1', 'TASK2' ) or taskname in ('TASK3',
  'TASK4' ) or taskname in ( 'TASK5', 'TASK6' ));
  
  Note the brackets in which the clause is enclosed.
  
  Because of this bug, the taskscheduler never gets updated,
  because the update fails. This failure happens while the
  overlapping tasks are running in the system. This could be why
  we are seeing multiple tasks running as the servers don't see
  the updated server timestamp and try to start it. None of the
  servers are able to update the servertimestamp because of the
  update fail.
  
  
  ENVIRONMENT (SYSTEM INFO):
  IBM Maximo Asset Management 7.6.0.5
  
  11.2.0.4.0 - 64bit Production With the Partitioning option)
  
  LOCAL FIX:
  None
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * When using multiple JVMs with more then 900 crontasks        *
  * scheduled.                                                   *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * Scheduled Job runs from Multiple JVMs when it is only        *
  * supposed to run one time.                                    *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * Upgrade to latest release                                    *
  ****************************************************************
  Steps to Reproduce:
  Run the same memory intensive escalation instance multiple,
  overlapping times.
  
  CURRENT ERRONEOUS RESULT:
  All of the jobs fail.
  
  EXPECTED RESULT:
  Single jobs runs to successful completion.
  
  ADDITIONAL INFORMATION
  The reason for this issue is the overlapping cron jobs. When
  such overlap happens, you would have multiple server entries for
  the same task names. The update query constructuted to update
  the taskschedule jobs looks like (Note that I have simplified
  the query to illustrate the issue)
  
  update taskscheduler set servertimestamp = TIMESTAMP ,
  servername = 'SERVER1' where servername = 'SERVER1' and taskname
  in ( 'TASK1', 'TASK2' ) or taskname in ('TASK3', 'TASK4' ) or
  taskname in ( 'TASK5', 'TASK6' );
  
  which when someone casually looks at it, can easily ignore and
  not see an issue. The problem is, becasue of the overlapping
  tasks running on multiple servers, the above query yields
  records for other servers and not just for SERVER1.
  
  The correct query should have been
  
  update taskscheduler set servertimestamp = TIMESTAMP ,
  servername = 'SERVER1' where servername = 'SERVER1' and
  (taskname in ( 'TASK1', 'TASK2' ) or taskname in ('TASK3',
  'TASK4' ) or taskname in ( 'TASK5', 'TASK6' ));
  
  Note the brackets in which the clause is enclosed.
  
  Because of this bug, the taskscheduler never gets updated,
  because the update fails. This failure happens while the
  overlapping tasks are running in the system. This could be why
  we are seeing multiple tasks running as the servers don't see
  the updated server timestamp and try to start it. None of the
  servers are able to update the servertimestamp because of the
  update fail.
  

Problem conclusion

• Scheduled Job run from single JVM when it is only supposed to
  run one time.
  
  The fix for this APAR is contained in the following maintenance
  package:
  	 | release\fix pack | Fix Pack Release 7.6.0.8 TPAE
  

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:

Modules/Macros

• MAXIMO
  

Fix information

• Fixed component name

  MAXIMO SYSTEMS

• Fixed component ID

  5724R46AV

Applicable component levels

• R760 PSY

     UP