OPS/MVS interfaces callable by z/OSMF
Article ID: 274973
Updated On:
Products
Issue/Introduction
We are automating various software deployments and have a query regarding the availability of interfaces supported by OPS/MVS that could be utilized by a product such as z/OSMF. We can envisage a use case where we could use a product such as z/OSMF to call OPS/MVS to perform some automation, then have OPS/MVS pass back the result to z/OSMF.
We can see various scenarios where we would use z/OSMF work flows to upgrade software running on z/OS. As part of those upgrades we may need to stop and start tasks that are managed using SSM and therefore the work flow would need to be able to drive that. Alternatively we may want to disable some OPS/MVS automation that monitors a set of tasks while we update related software and recycle tasks that run that software. There could also be a situation where we use OPS/MVS automation routines to perform health checks and following a software upgrade we may want to trigger those automation routines and retrieve the result, before continuing with the work flow.
Please could you help clarify whether OPS/MVS supports interfaces that could be utilized by a product such as z/OSMF to enable this type of interaction?
Environment
Release : 14.0
Resolution
Both OPS/MVS and z/OSMF can utilize the REST API, however an API is not necessary to accomplish the desired result. A much simpler and more flexible approach follows:
1. All scenarios mentioned can be accomplished via OPS/REXX programs.
2. OPS/REXX programs can be executed from JCL batch jobs.
3. JCL jobs that call an OPS/REXX program can be submitted by a z/OSMF workflow step.
4. Any OPS automation command output or error messages gathered in the rexx code can be passed back to the job's JESMSGLG via SAY statements, where they can be easily assessed in the JESMSGLG tab for the workflow step.
5. The return code set by the rexx execution can be passed back to the job, which in turn will set the return/condition code for the job.
6. A zero return code reported by the z/OSMF job step will allow the z/OSMF workflow to continue to the next step, whereas a non-zero return code will fail the workflow step and the workflow will stop.
As a simple example, let's say there is a z/OSMF workflow step that needs to check that 4 global variables are defined to OPS/MVS and values are present.
The following is the CHKGLBL OPS/REXX program that will check that the global variables are defined and report their values:
retcode = 0 <<<< Set an initial RC of 0
do i=1 to 4
varname = 'GLOBAL.TEST.'i
value = opsvalue(''varname'','0')
if value = 0 then
do
retcode = 8 <<<< Set RC 8 if a global is missing
say varname 'DOES NOT EXIST!' <<<< SAY statements can be used to pass any output back to the job's JESMSGLG
end
else
say varname '=' value <<<<
end
return retcode <<<< Send the return code back to the job
The following is the JCL defined in the workflow step to call the CHKGLBL rexx referenced above:
<jobcard>
//STEP1 EXEC PGM=OI,PARM=CHKGLBL <<<< This executes the rexx member CHKGLBL
//SYSIN DD DUMMY
//STEPLIB DD DISP=SHR,DSN=hlq.CCLXLOAD <<<< Steplib to the OPS loadlib
//SYSEXEC DD DISP=SHR,DSN=your.rexx.library <<<< REXX library where the CHKGLBL rexx resides
//OPS$OPSS DD DUMMY <<<< Set the SSID of the OPS/MVS region to execute the rexx, in this case OPSS
//SYSPRINT DD SYSOUT=*
//SYSTSIN DD DUMMY
If all the globals are present and accounted for, the following will be returned to the job's JESMSGLG which can be accessed via the workflow step's JESMSGLG tab:
19.43.28 JOB06618 ---- FRIDAY, 06 OCT 2023 ----
19.43.28 JOB06618 TSS7000I MYUSERID Last-Used 06 Oct 23 15:40
19.43.28 JOB06618 TSS7001I Count=00383 Mode=Fail Locktime=No
19.43.28 JOB06618 $HASP373 TESTJOB STARTED - INIT 1 - CL
19.43.28 JOB06618 IEF403I TESTJOB - STARTED - TIME=19.43.28
19.43.28 JOB06618 +OPS0996I GLOBAL.TEST.1 = FRED <<<< Here is the output passed from the rexx execution
19.43.28 JOB06618 +OPS0996I GLOBAL.TEST.2 = WILMA <<<<
19.43.28 JOB06618 +OPS0996I GLOBAL.TEST.3 = BARNEY <<<<
19.43.28 JOB06618 +OPS0996I GLOBAL.TEST.4 = BETTY <<<<
19.43.28 JOB06618 IEF404I TESTJOB - ENDED - TIME=19.43.28
19.43.28 JOB06618 $HASP395 TESTJOB ENDED - RC=0000 <<<< The RC 0 passed back from the rexx will allow the workflow to continue
Now, what if a global is missing? In that case, the following will be returned to the job's JESMSGLG:
19.54.22 JOB06619 ---- FRIDAY, 06 OCT 2023 ----
19.54.22 JOB06619 TSS7000I MYUSERID Last-Used 06 Oct 23 15:40
19.54.22 JOB06619 TSS7001I Count=00383 Mode=Fail Locktime=No
19.54.22 JOB06619 $HASP373 TESTJOB STARTED - INIT 1 - CL
19.54.22 JOB06619 IEF403I TESTJOB - STARTED - TIME=19.54.22
19.54.22 JOB06619 +OPS0996I GLOBAL.TEST.1 = FRED
19.54.22 JOB06619 +OPS0996I GLOBAL.TEST.2 DOES NOT EXIST! <<<< There's a problem
19.54.22 JOB06619 +OPS0996I GLOBAL.TEST.3 = BARNEY
19.54.22 JOB06619 +OPS0996I GLOBAL.TEST.4 = BETTY
19.54.22 JOB06619 IEF404I TESTJOB - ENDED - TIME=19.54.22
19.54.22 JOB06619 $HASP395 TESTJOB ENDED - RC=0008 <<<< The RC 8 passed back from the rexx will fail the workflow step and stop the workflow.
The above is a simple example for the sake of brevity, but if necessary, complex processing and reporting can be implemented within any OPS/REXX program called via a JCL step in the z/OSMF workflow, allowing control of virtually every aspect of OPS/MVS. Documentation of OPS/REXX commands/functions specific to the example scenarios that were mentioned follow:
"we may need to stop and start tasks that are managed using SSM and therefore the work flow would need to be able to drive that."
"we may want to disable some OPS/MVS automation that monitors a set of tasks while we update related software and recycle tasks that run that software."
"we use OPS/MVS automation routines to perform health checks and following a software upgrade we may want to trigger those automation routines and retrieve the result, before continuing with the work flow."
If the "automation routines" are already in the form of OPS/REXX programs, then they too can be called via a z/OSMF JCL workflow step.
Feedback
