Using Operating System Scheduling on z/OS Systems

Limitations of z/OS Operating System Scheduling Servers

On z/OS operating system scheduling servers, SAS 9.3 supports the scheduling of flows that have a single time event as a trigger. The jobs in a flow are submitted to the Job Entry Subsystem (JES) for execution in an asynchronous manner.

How Operating System Scheduling Works on z/OS Systems

When you schedule a flow, Schedule Manager creates the following files:

a REXX program called RunFlow_yyymmddhhmmsssss.rex, which submits each batch job to run on the requested schedule.
a file called jobname.status, which contains information about the status of the job.
the following files, each of which invokes the REXX program to log the job status:
- logstart_jobname, which is executed as the first step of each job
- logend_jobname, which is executed as the last step of the job if all previous steps have a return code of 0
- logfail_jobname, which is executed as the last step of the job if any of the previous steps has a return code other than 0
flowname.cmd, which contains the statements that execute the program for the flow. This file is passed to the at command to be executed when the time event trigger occurs.
flowname.lck, which serializes the execution of the REXX exec by simultaneously executing jobs.
flowname.log, which contains the log information for the flow.
flowname.status, which contains the current status of the flow.

The files are written to a subdirectory called username/flowdefname in the Control Directory that was specified in the operating system scheduling server definition, where username is the identity that scheduled the flow.

Manually Submitting a Flow for Execution on a z/OS Operating System Scheduling Server

Note: On z/OS, unexpected results can occur if you manually submit a flow without having selected Manually Schedule when you scheduled the flow.

If you need to manually submit a flow that has been scheduled on a z/OS operating system scheduling server, follow these steps:

Change to the directory where the flow's files are located (/outdir/username/flowdefname), where outdir is the value of the Control Directory attribute in the operating system scheduling server definition.
At the USS prompt, type a command with the following syntax:
```
RunFlow_yyymmddhhmmsssss.rex <runonce|start>
```
Specify runonce if you want to run the flow just once, or specify start if you want to resume a recurrence that was specified for the flow.

Canceling a Scheduled Flow on a z/OS Operating System Scheduling Server

If you need to cancel a flow that has been scheduled on a z/OS operating system scheduling server, follow these steps:

At the USS prompt, type at -l.

Note: Type a lowercase L, not the numeral 1.

A list of jobs appears, as in the following example:

1116622800.a   Fri May 20 17:00:00 2005
1116734400.a   Sun May 22 00:00:00 2005
1116820800.a   Mon May 23 00:00:00 2005
1116620801.a   Mon May 23 00:00:00 2005
1117029000.a   Wed May 25 09:50:00 2005

For each job that you want to cancel, type the following command at the USS prompt: at -r at-jobid, where jobid is the first part of each line of the list of jobs (for example, at -r 1116622800.a).