4.5.11 Rerunning a job or jobnet

(1) Rerun the root jobnet

You can choose one of the following rerun methods when you rerun a root jobnet.

From abnormally ended job

This method reruns from an abnormally ended job. Even if a job that terminated abnormally is preceded by a job that ended normally, the job that ended normally is not rerun.

If the rerun job ends normally, succeeding jobs are executed and processing continues.

When a nested jobnet has ended abnormally, you can rerun from the specific job in the nested jobnet that ended abnormally.

From after abnormally ended job

This method reruns from the job that follows the abnormally ended job.

When a nested jobnet has ended abnormally, you can rerun from the job that follows the job that ended abnormally in the nested jobnet.

However, among the units connected by a relation line, if the status of the first unit that ended abnormally is Not executed + Ended or Skipped so not exe., those units will also rerun the job.

From abnormally ended jobnet

This method reruns from the jobnets directly under which jobs that terminated abnormally are included.

Even if the jobs directly under a nested jobnet include both jobs that terminated abnormally and jobs that ended normally, all jobs directly under the nested jobnet are rerun along with their succeeding units.

If there are any jobs that terminated abnormally directly under the root jobnet, the entire root jobnet is rerun.

From top of jobnet

This method reruns from the top of the root jobnet.

Only jobs that ended with warning

This method reruns only jobs that ended with a warning. When a nested jobnet is defined, you can rerun the jobs that ended with a warning in the nested jobnet.

In addition, rerun jobs that have ended once with a warning and then terminated abnormally when they were rerun.

To Page Top

(2) Rerun a nested jobnet or job

You can choose from the following methods to rerun a nested jobnet or job.

From the specified jobnet or job

This method reruns from the specified jobnet or job. If the rerun jobnet or job ends normally, succeeding jobs are executed and processing continues.

From the job after the specified jobnet or job

This method reruns from the job that follows the specified job or nested jobnet.

Only the specified jobnet or job

This method reruns only the specified job or nested jobnet.

Rerunning a jobnet or job results in an error in the following cases:

• The upper-level root jobnet of the jobnet or job to be re-executed is placed in the Being held or Waiting for the start time status.

• The jobnet or job to be re-executed is in the Now Running, Running + Abend, Running + Warning, Wait for prev. to end, Wait for start time, Being held status, Not sched. to exe., Now queuing, or Waiting to execute.

• The jobnet or job to be re-executed has the Recovery attribute, and the preceding jobnet or job has ended normally or with a warning.

• The jobnet or job to be re-executed is a judgment job or a dependent unit of a judgment job.

  Note that dependent units of judgment jobs are re-executable if yes is set for the RERUNSUBORDINATE environment setting parameter. For details about whether to enable re-execution of dependent units, see (4) Re-executing the dependent unit of a judgment job.

• The upper-level jobnet of the jobnet or job to be re-executed does not have a schedule.

In the following cases, attempting to execute a job or jobnet does not result in an error but they will not execute:

• When execution of the job to be re-executed is canceled

• When the jobnet that you attempted to re-execute does not have a schedule, even though its upper-level jobnet does have a schedule

To Page Top

(3) The job preceding a judgment job ended abnormally

When the job preceding a judgment job ends abnormally, you can rerun that preceding job as follows.

• Rerun the root jobnet

  • From abnormally ended job

  • From after abnormally ended job^#1

  • From abnormally ended jobnet

  • From top

  • Only jobs that ended with warning^#2

• Rerun a nested jobnet or job

  • From this unit (excludes dependent units)

  • After this unit (excludes dependent units)^#1

  • Only this unit (excludes dependent units)

#1

Behavior when rerunning from the next unit whose preceding job was ended abnormally:

• When a job preceding a judgment job ends abnormally, if you rerun the jobnet from the job subsequent to the job that terminated abnormally, the status of the job that terminated abnormally changes to Ended normally, and the execution of processing begins from the next judgment job.

• When the judgment job makes a judgment using the return code of the preceding job, it uses the return code from when the preceding job ended abnormally.

#2

This applies to the preceding job that ended abnormally because it was re- executed after it ended with a warning.

To Page Top

(4) Re-executing the dependent unit of a judgment job

The execution of the dependent unit of a judgment job depends on the execution result of the judgment job. If the dependent unit associated with a judgment job terminates abnormally, whether the dependent unit can be re-executed and how to re-execute it depend on the setting of the RERUNSUBORDINATE environment setting parameter. For details about the parameter, see the appropriate manual section for the platform you are using. If you are using Windows, see 6.2.23 Settings for re-executing the dependent units in the JP1/Automatic Job Management System 3 Configuration Guide. If you are using UNIX, see 15.2.21 Settings for re-executing the dependent units in the JP1/Automatic Job Management System 3 Configuration Guide.

If re-execution of dependent units is disabled (environment setting parameter RERUNSUBORDINATE = no):

Unlike ordinary jobs (such as PC jobs and Unix jobs), dependent units cannot be re-executed. Re-execution from the judgment job that precedes a dependent unit is also not possible.

If a dependent unit terminates abnormally, re-execute the jobnet that contains the dependent unit.

If re-execution of dependent units is enabled (environment setting parameter RERUNSUBORDINATE = yes):

Whether dependent units can be re-executed depends on the unit type and the status of the dependent units.

The dependent unit of a judgment job can be re-executed when all of the following conditions exist:

• The version of JP1/AJS3 - Manager is 10-00 or later.

• The judgment job evaluates to true.

• The judgment job is in Ended normally status.

• In the following table, Y indicates the unit type of a dependent unit that can be re-executed.

  Table 4‒6: Unit types of dependent units that can be re-executed

  No.	Unit type	Unit defined as a dependent unit	Unit defined in a dependent jobnet
  1	Job group	--	--
  2	Root jobnet	--	--
  3	Nested jobnet	Y	Y
  4	Root remote jobnet	--	--
  5	Nested remote jobnet	Y	Y
  6	Manager job group	--	--
  7	Manager jobnet	--	--
  8	Planning group	--	--
  9	Standard job	Y	Y
  10	Jobnet connector	--	Y
  11	OR job	--	Y
  12	Judgment job	--	N
  13	Event job	Y	Y
  14	Action job	Y	Y
  15	Custom job	Y	Y
  16	Passing information setting job	Y	Y
  17	HTTP connection job	Y	Y
  18	Custom event job	Y	Y

  Legend:

  Y : The unit can be re-executed (if all necessary conditions are satisfied).

  N : The unit cannot be re-executed.

  -- : The unit cannot be defined as a dependent unit.

• In the following table, Y indicates a dependent unit in the indicated can be re-executed.

  Table 4‒7: Statuses and types of dependent units that can be re-executed

  No.	Status	Type of dependent unit
  		Jobnet	Job	Jobnet connector
  1	Not sched. to exe.	N	N	N
  2	Wait for start time	N	--	--
  3	Wait for prev. to end	N	N	N
  4	Being held	N	N	--
  5	Waiting to execute	--	N	--
  6	Now queuing	--	N	--
  7	Not executed + Ended	Y	Y	Y
  8	unexec-W	Y	Y	--
  9	Bypassed	Y	Y	Y
  10	Now running	N	N	N
  11	Running + Abend	N	--	N
  12	Running + Warning	N	--	N
  13	Ended normally	Y	Y	Y
  14	Normal end + False	--	--	--
  15	Ended with warning	Y	Y	Y
  16	Ended abnormally	Y	Y	Y
  17	abnormal-WR	Y	Y	--
  18	Skipped so not exe.	N	--	--
  19	Invalid exe. seq.	Y	--	--
  20	Interrupted	Y	--	--
  21	Killed	Y	Y	Y
  22	kill-WR	Y	Y	--
  23	Failed to start	--	Y	--
  24	fail-WR	--	Y	--
  25	Unknown end status	--	Y	Y
  26	unknown-WR	--	Y	--
  27	Shutdown	N	N	N
  28	Wait for start cond.	--	--	--
  29	Now monitoring	--	--	--
  30	Unmonitored + Ended	--	--	--
  31	Monitor terminated	--	--	--
  32	Interrupted monitoring	--	--	--
  33	Monitor-end normal	--	--	--

  Legend:

  Y : The unit can be re-executed.

  N : The unit cannot be re-executed.

  -- : The unit cannot have the indicated status.

To Page Top

(5) Notes about rerunning jobnets or jobs

Take note of the following when you rerun a jobnet or job.

• Processes from before the rerun have still not ended

  When you rerun a jobnet from a specific job, the rerun job may catch up with a job that started before the rerun and still has active processes. In this case, start the rerun job after the preceding job ends.

• Result information for a job before the rerun

  When you rerun a job, the result information for the job (execution start time, execution end time, status, code) is updated to reflect the results of the rerun. If the Append check box is not selected, the information in the standard output file and the standard error output file is updated to reflect the results of the rerun. However, messages that JP1/AJS3 outputs to the standard error output file might be added even if the Append check box is not selected.

• Changing the status of a jobnet or job that precedes the rerun jobnet

  If the status of the preceding jobnet or job was Ended abnormally, Interrupted, Invalid exe. seq., Killed, Failed to start, Unknown end status, Skipped so not exe., or Not executed + Ended, on rerunning the status is forcibly changed to Ended normally, Ended with warning, or Bypassed. The following table details these changes.

  Table 4‒8: Status change of units

  No.	Status of preceding unit	Does the status apply?#1			After status of preceding unit is changed
  		Jobnet		Job	Set skipped jobs to warning?
  		Root	Nested		Yes	No
  1	Not sched. to exe.	Y	Y	Y	Not sched. to exe.	Not sched. to exe.
  2	Waiting for start time	Y	Y	N	Waiting for start time	Waiting for start time
  3	Waiting for prev. to end	N	Y	Y	Waiting for prev. to end	Waiting for prev. to end
  4	Being held	Y	Y	Y	Being held	Being held
  5	Waiting to execute	N	N	Y	Waiting to execute	Waiting to execute
  6	Now queuing	N	N	Y	Now queuing	Now queuing
  7	Not executed + Ended [-W]	N	Y	Y	Ended with warning	Bypassed
  8	Bypassed	N	Y	Y	Bypassed	Bypassed
  9	Now running	Y	Y	Y	Now running	Now running
  10	Running + Warning	Y	Y	N	Running + Warning	Running + Warning
  11	Running + Abend	Y	Y	N	Running + Abend	Running + Abend
  12	Ended normally	Y	Y	Y	Ended normally	Ended normally
  13	Ended with warning	Y	Y	Y	Ended with warning	Ended with warning
  14	Ended abnormally [-WR]	Y	Y	Y	Ended with warning	Ended normally
  15	Skipped so not exe.	Y	Y	N	Ended with warning	Bypassed
  16	Invalid exe. seq.	Y	Y	N	Ended with warning	Ended normally
  17	Interrupted	Y	Y	N	V5: Ended with warning #2 V6: Interrupted #2	V5: Ended normally #2 V6: Interrupted #2
  18	Killed [-WR]	Y	Y	Y	Ended with warning	Ended normally
  19	Failed to start [-WR]	N	N	Y	Ended with warning	Ended normally
  20	Unknown end status [-WR]	N	N	Y	Ended with warning	Ended normally
  21	Shutdown	Y	Y	Y	N/P	N/P
  22	Normal end + False	N	N	Y	Normal end + False	Normal end + False
  23	Wait for start cond.	Y	Y#3	N	N/P	N/P
  24	Now monitoring	Y	Y#3	N	N/P	N/P
  25	Unmonitored + Ended	Y	Y#3	N	N/P	N/P
  26	Monitor terminated	Y	Y#3	N	N/P	N/P
  27	Interrupted monitoring	Y	Y#3	Y#3	N/P	N/P
  28	Monitor-end normal	Y	Y#3	N	N/P	N/P
  29	Start delay	Y	Y	N	#4	#4
  30	End delay	Y	Y	Y	#4	#4
  31	Nested jobnet delayed start	Y	Y	N	#4	#4
  32	Nested jobnet delayed end	Y	Y	N	#4	#4

  Legend:

  Root: Root jobnet

  Nested: Nested jobnet

  Y: Applies

  N: Does not apply

  N/P: Rerun is not possible (an error message is output)

  Note

  Numbers 1 to 6 and 9 to 11 indicate the status when the preceding unit that was terminated is rerun.

  #1

  Indicates whether each type of unit can enter the indicated status.

  #2

  Partial rerunning of an interrupted unit depends on the setting (V5 or V6) of the environment parameter VRSHIFT_INTRERUN. It is set to V6 by default. During a new installation and setup of JP1/AJS3, V5 is set. The table shows the status for each setting.

  #3

  Since this status only applies to a start condition (.CONDITION) or an event icon in a start condition, the unit cannot be a preceding unit.

  #4

  Delay is additional information relating to the unit status. Therefore, the delay status does not change even if the succeeding unit is rerun.

• When rerunning the succeeding unit of a unit that has terminated abnormally

  If a unit has terminated abnormally, all the succeeding units other than recovery units move to the Not executed + Ended status. The status is changed unit by unit in the order in which the units are linked with relation lines. Accordingly, when you rerun the succeeding unit of the abnormally terminated unit, the status of the succeeding unit might not move from the Waiting for prev. to end status to the Not executed + Ended status, depending on the timing. Before you attempt to rerun the succeeding unit of the abnormally terminated unit, wait until all the succeeding units are placed in the Not executed + Ended status. You must be especially careful when you rerun the succeeding unit from a recovery unit.

  If you rerun the succeeding unit of an abnormally terminated unit when the succeeding unit is in the Wait for prev. to end status, the following problems might occur:

  • An attempt to rerun the unit fails

    This problem occurs when the unit to be rerun is still in the Wait for prev. to end status.

    Figure 4‒37: When the unit in the Wait for prev. to end status is rerun

    Job C is in the Wait for prev. to end status and therefore cannot be rerun. Wait until Job C moves to the Not executed + Ended status, and then rerun it.

  • An attempt is made to rerun only the specified unit, but succeeding units are also run

    This problem occurs when you rerun a unit in the Not executed + Ended status while another unit is in the Wait for prev. to end status, resulting in no preceding units in the Ended abnormally status.

    Figure 4‒38: When the succeeding job has a unit in the Wait for prev. to end status

    When Job C in the Not executed + Ended status is rerun, Job C moves to the Wait for prev. to end status, Job A moves to the Ended normally status, and Job B moves to the Bypassed status. Because none of the preceding units for Job D are in the Ended abnormally status, Job D does not move to the Not executed + Ended status, and will therefore run after Job C is rerun.

  • If there is a branching unit between a terminated abnormally unit and the specified unit, the branching unit is run.

    This problem occurs if the following conditions are met: 1) a branching unit in Wait for prev. to end status remains; 2) units in Not executed + Ended status are rerun; 3) no units preceding the branching unit in Wait for prev. to end status are in Ended abnormally status.

    Figure 4‒39: When the branching job has a unit in the Wait for prev. to end status

    When re-execution starts with Job C in Not executed + Ended status, the status of jobs changes as follows: the status of Jobs C and D changes to Wait for prev. to end, the status of Job A changes to Ended normally, and the status of Job B changes to Bypassed. Because none of the preceding units for Job E are in Ended abnormally status, Job E is run without being placed in Not executed + Ended status.

  • If there are multiple terminated abnormally preceding units, the other preceding units between the terminated abnormally ones and the unit specified for re-execution are run.

    This problem occurs if the following conditions are met: 1) a unit in Wait for prev. to end status remains between units in Ended abnormally status and the unit that is in Not executed + Ended status and is specified for re-execution; 2) the unit in Not executed + Ended status in 1) is rerun; 3) no units preceding the unit in Wait for prev. to end status in 1) are in Ended abnormally status.

    Figure 4‒40: When the preceding job has a unit in the Wait for prev. to end status

    When re-execution starts with Job C in Not executed + Ended status, the status of jobs changes as follows: the status of Jobs C and D changes to Wait for prev. to end, the status of Job A and Job E changes to Ended normally, and the status of Job B changes to Bypassed. Because none of the preceding units for Job F are in Ended abnormally status, Job F is run without being placed in Not executed + Ended status.

• When concurrent execution is disabled

  If you rerun a jobnet for which concurrent execution is disabled while another generation of the jobnet is running, the status of the re-executed generation changes to Waiting for start time and it is executed when the other generation finishes.

  In the same circumstances but with concurrent execution enabled, the jobnet is re-executed immediately.

• When Schedule skip is set as the schedule option

  Execution schedules are not skipped when a jobnet whose schedule option is set to Schedule skip is rerun and the scheduled start time of the next run is reached during the re-execution. In this case, the execution schedule, which was not skipped, is executed in accordance with the settings for concurrent executions. The Schedule skip option applies only when the next start time arrives and the preceding generation was executed at its scheduled start time. For details about skipping a schedule, see 3.3.3(2) Schedule option.

• When rerunning a root jobnet or preceding unit as the result of a judgment job

  Take care when rerunning a root jobnet or preceding unit as the result of a judgment job. For details, see 2.4.3 Dynamically changing a process depending on the result of a preceding job (example of defining a jobnet that uses a judgment job) in the JP1/Automatic Job Management System 3 System Design (Work Tasks) Guide.

To Page Top

(6) Notes on resuming processing while a recovery unit is running

If a job has terminated abnormally, JP1/AJS3 starts a sequence of recovery units. However, if you resume the processing from the job after the abnormally terminated one, the rest of the recovery units will be placed in the Not executed + Ended status and will not be executed. In this case, resume processing after all recovery processes have been performed.

In addition, if you resume units in a way that causes a currently running recovery jobnet to be resumed, such as by specifying a preceding unit that terminated abnormally and resuming that unit with the option After this unit specified, the following might occur:

When the preceding unit is resumed and its status becomes something other than Ended abnormally, the statuses of the units under the recovery jobnet become Not executed + Ended, and execution of the recovery jobnet is interrupted.

To resume the succeeding unit of the preceding unit that terminated abnormally without interrupting the execution of the recovery jobnet, resume units so that the recovery jobnet is not resumed, such as by specifying and resuming the succeeding unit immediately after the preceding unit with the From this unit option specified.

When you rerun the succeeding unit by using the ajsrerun command in a recovery unit, if you specify the -rr, -rn, and -ri options for the command, re-execution can be retried when the unit to be rerun has not ended. For details about the ajsrerun command, see ajsrerun in 3. Commands Used for Normal Operations in the manual JP1/Automatic Job Management System 3 Command Reference.

Note that retry is performed only when the unit specified in the ajsrerun command is not ended as shown in Figure 4-37. In cases shown as Figure 4-38, Figure 4-39, and Figure 4-40, the recovery unit cannot retry re-execution and wait until the other units are ended. For a definition example of rerunning the succeeding unit by using the ajsrerun command in a recovery unit, see 2.4.6 Executing a specific process when a job ends abnormally (example of defining a jobnet that uses a recovery unit) in the JP1/Automatic Job Management System 3 System Design (Work Tasks) Guide.

If you want to resume the subsequent jobs even when an error is detected, consider whether you can use a judgment job to perform recovery with a dependent unit when the termination code of the judgment job has exceeded a threshold.

To Page Top

(7) Notes on re-executing a jobnet with start conditions

When monitoring of start conditions for a root jobnet is completed and the jobnet status changes from Now monitoring to any of the following statuses, the status of the lower-level units becomes Not executed + Ended.

• Unmonitored + Ended

• Monitor terminated

• Monitor-end normal

Because the Not executed + Ended status is treated as an abnormal termination, be careful when you re-execute a generation of a jobnet that has any of the following statuses.

To Page Top