Setting NON Zero return codes in ESP Batch

[Lizette]

I have developed a process to set a non zero return code for ESP batch functions.   I am currently using this in the History Archive process when we close the HIST1 , Copy the History Records and Archive Records, and open HIST1.

 

It was suggested I put this out to the community in-case anyone else is interested in using it.

 

The Psuedo code is

ESPBatch

 Write SYSPRINT to Temp file

IEBGENER

  Print Temp file back to SYSOUT

IKJEFT01 

    Use Rexx to scan for specific messages and set the return code.

 

If this is of interest I will be happy to post the three steps plus the REXX.

 

Here is the REXX code.  It is very simple and is attached.

 

 

This was the only way I could find to set a return code when an ESP Batch step got an error.

 

The rewrite of the ESP History File Archive went from 6 steps to 14.

 

Step Name	Restart? Y/N	Step Program	Description
STEP01	Y	ESP	Close the HIST1 file for processing
STEP02	N	IEBGENER	Print the ESP Batch file output
STEP03	N	IKJEFT01	Parse the ESP Batch output and set error code
STEP04	Y	IDCAMS	Backup the HISTFILE to a temp home
STEP05	Y	ESP	Copy the Archive Records and Reload records
STEP06	N	IEBGENER	Print the ESP Batch file output
STEP07	N	IKJEFT01	Parse the ESP Batch output and set error code
STEP08	Y	IDCAMS	Delete/Define the ESP Histfile
STEP09	Y	SORT	Sort the records to be reloaded into the HISTFILE
STEP10	Y	IDCAMS	Load the records into the HISTFILE
STEP11	Y	ESP	OPEN the HIST1 file
STEP12	N	IEBGENER	Print the ESP Batch file output
STEP13	N	IKJEFT01	Parse the ESP Batch output and set error code
STEP14	Y	IDCAMS	List the HISTFILE for later review for trending for size

 

Lizette

Attachment(s)

ESP Rexx to parse messages.txt.zip   1 KB 1 version

[Lucy Zhang]

Hi Lizette,

 

Thank you for sharing.  I changed it from a question to a discussion.

 

Do you use BATCHRC in ESP init parm? Wonder if it can help to show the proper return code.

BATCHRC MAXSEV|LEGACY

MAXSEV causes the return code to reflect the most severe error detected. LEGACY keeps the original behavior. The original behavior is preserved unless MAXSEV is explicitly requested. The setting remains active until changed by BATCHRC command or until IPL. The new command applies to master and proxy subsystems.

 

Note: It seems the above is not in the ESP related manuals. I will seek to add it.

 

Lucy

[Lizette]

Thank you Lucy.

 

The only reference I could find is in this manual Dynamic Agent Configuration Enhancement in section SAVEAGDF Command: Save Agent Definitions

 

But it does not describe this parameter or how it is used.

This is the example:

//NEWGEN EXEC PGM=IEFBR14
//AGENTDEF DD DSN=GL.PROD.TEMPAGDF(+1),DISP=(,CATLG,DELETE),
// UNIT=SYSDA,VOL=SER=ESPVOL,SPACE=(TRK,1),
// DCB=(DSORG=PS,RECFM=FB,BLKSIZE=23440,LRECL=80)
//*
//ESP EXEC PGM=ESP,PARM='SUBSYS(ESPM)',COND=(0,LT)
//STEPLIB DD DISP=SHR,DSN=GL.PROD.LOADLIB
//SYSPRINT DD SYSOUT=*
//SYSIN DD *
OPER BATCHRC MAXSEV
SAVEAGDF 'GL.PROD.TEMPAGDF(0)'

 

Lizette

[Lucy Zhang]

Our doc team will add BATCHRC related information in the ESP online manual.

 

Sorry about that.

 

Lucy

[Michael Bieganski]

Lucy, while awaiting the addition of BATCHRC to the online manuals, could you please elaborate a bit on what it all entails?

Must a particular ptf be applied before this can be done in pagemode?

As Lizette points out, there are some ESP batch cmds that simply always end in 0, and you must eyeball the sysprint to see if there was any squawking......How could BATCHRC help with that assuming its a one step job that returns 0 right now? 

Currently, if we do an LI on a job, it appears to show just the last rc from a multi-step job (often a 0 even if prior steps blow up six ways to Sunday).  If we issued BATCHRC MAXSEV in pagemode on the master and each proxy, would that change to start showing the highest RC as opposed to the last?

[Lucy Zhang]

Hi Michael,

 

It should be on the base of version 11.4.

 

If the ESP command returns zero, then this parm won't help. It helps in the scenario that multiple ESP commands are entered in same ESP batch step, and some (not the last) return non-zero return code.

 

As the Tuesday tip in the link below, this BATCHRC only affects ESP batch step (not even the ESP batch job) and ESP report step (like SADGEN), it won't affect non-ESP batch jobs:

https://communities.ca.com/message/100614154

 

It won't help on the LI command.

 

Thanks,

 

Lucy

[Michael Bieganski]

Thanks Lucy

[Michael Bieganski]

very nice Lizette.  Would you please post the '3 steps'?   This should prove valuable for esp cmds that do not bow to the batchrc cmd. (for example I did a batch  resdef b- list   and it came back with a cc 0 and 

ESP1905I No Resources found matching B-

Adding the batchrc maxsev did not change anything:

ESP VER 11.4 BATCH INTERFACE

OPER BATCHRC MAXSEV
ESP2791I Batch RC mode is MAXSEV
RESDEF B- LIST
ESP1905I No Resources found matching B-

cc was still 0, so in cases such as this, your eloquent rexx procedure might be the only solution currently.

thanks

[Lizette]

Here is the JCL

 

//STEP01 EXEC PGM=ESP,PARM='SUBSYS(ESP)'
//SYSABEND DD SYSOUT=U,FREE=CLOSE,SPIN=UNALLOC
//SYSUDUMP DD SYSOUT=U,FREE=CLOSE,SPIN=UNALLOC
//*
//SYSPRINT DD DISP=(,PASS),UNIT=SYSDA,DSN=&&ESPTEMP,
// LRECL=135,RECFM=FB,BLKSIZE=0,SPACE=(CYL,(5,100),RLSE)
//*
//SYSIN DD *
OPER TRACKING NOSTORE
OPER HISTFILE HIST1 CLOSE
/*
//*
//* -----------------------------------------------------------------
//*
//* STEP02: THE FOLLOWING IEBGENER WILL PRINT THE OUTPUT FROM
//* THE ESP BATCH STEP TO SYSOUT. SO THE INFORMATION IS
//* MAINTAINED WITHIN THE JOB STREAM.
//*
//* -----------------------------------------------------------------
//*
//STEP02 EXEC PGM=IEBGENER DO NOT CONDITION CODE THIS STEP
//SYSPRINT DD SYSOUT=*
//*
//SYSUT1 DD DISP=(OLD,PASS),DSN=&&ESPTEMP
//SYSUT2 DD SYSOUT=*
//*
//SYSIN DD DUMMY
//*
//* -----------------------------------------------------------------
//*
//* STEP03: PARSE MESSAGES. SEE IF THERE IS A MESSAGE IN
//* THE ESP BATCH STEP THAT NEEDS A DIFFERENT RC
//*
//* -----------------------------------------------------------------
//*
//STEP03 EXEC PGM=IKJEFT01,DYNAMNBR=100,PARM='%RXESPMSG'
//*
//SYSPROC DD DISP=SHR,DSN=SYS1.CAI.ESP.REXX
//*
//* -----------------------------------------------------------------
//*
//* THE ESPRPTIN IS THE TEMP FILE FROM THE ESP BATCH STEP.
//* THE REXX RXESPMSG WILL BE USED TO PARSE THROUGH THE OUPUT
//* AND IF ANY MESSAGE ARE IN THE OUTPUT THAT SHOULD CAUSE
//* THIS PROCESS TO FAIL, IT WILL SET AN APPROPRIATE RC
//*
//* -----------------------------------------------------------------
//*
//ESPRPTIN DD DISP=(OLD,DELETE),DSN=&&ESPTEMP
//*
//SYSPRINT DD SYSOUT=*
//SYSPROC DD DISP=SHR,DSN=SYS1.CAI.ESP.REXX
//SYSTSPRT DD SYSOUT=*
//SYSTERM DD DUMMY
//SYSOUT DD SYSOUT=*
//*
//SYSTSIN DD DUMMY

 

 

Lizette

[Michael Bieganski]

thank you Lizette