JCL Interview Questions

JCL Interview Questions for Experienced Professionals

At this level, a person should not only be able to identify JCL syntax but also explain how to use it in real-world mainframe projects to maximise workloads and solve complex problems. Your responses must exhibit pragmatic understanding and application of concepts.

30. How do I use GDGs in JCL?

In an unordered list, I use Generation Data Groups (GDGs) to organise datasets that appear in chronological order (like daily input files). A GDG base can be constructed using IDCAMS, and each dataset can be described as a generation (+1, 0, -1). For example:

//STEP1 DD DSN=MY.GDG(+1),DISP=(NEW,CATLG,DELETE)

This sets up the next generation. If I wish to read the most recent one, I select (0). GDGs allow me to automate a daily/weekly job without the need to manually change dataset names.

31. How do I handle out-of-space errors (SB37, SD37, SE37)?

Based on my knowledge, I would state that these abends are caused when a dataset is deprived of primary or secondary space. I resolve these issues by:

• Increasing the primary and secondary space values contained within the SPACE parameter.
• Switching to a different UNIT or VOLUME with a larger capacity.
• Occasionally switching from TRACKS to CYLINDERS.

Example:

SPACE=(CYL,(50,10),RLSE)

This stops space abends by allocating additional cylinders.

32. What is the difference between system abend codes and user abend codes?

The difference between system abend codes and user abend codes is:

• Systems abends (Sxxx) are generated by the OS (i.e., S0C7 for a data exception, S0C4 for an invalid memory).
• User abends (Uxxx) are done by the program logic, usually through the use of the ABEND statement in COBOL or assembler.

As an example, when I encounter a U0016 abend, it often indicates that there has been a user-defined check that has failed.

33. How do I use the TIME parameter in JCL?

I can assign the TIME variable anywhere in the JOB or at the STEP level.

Example:

//STEP1 EXEC PGM=MYPROG,TIME=(1,30)

This scenario assigns STEP1 for a minute and thirty seconds. For an unlimited time, I use TIME=MAX. In production, I assign limits mainly to stop jobs from running uncontrollably and using excess resources.

34. What is the difference between HOLD and TYPRUN=HOLD in JCL?

Here is the difference between HOLD and TYPRUN=HOLD:

• HOLD on JOB statement: Offers the ability to submit a job in the input queue, and the job stays there until the operator decides to release it.
• TYPRUN=HOLD: I can use this in tests or when I need an operator to check the job before the system executes it.

I use this when I am trying to troubleshoot sensitive production jobs.

35. How do I use symbolic parameters in JCL?

Symbolic parameters apply to my PROCs as they can be made ‘flexible’ and defined inside the PROC using &, and values supplied in the calling JCL.

In PROC

//STEP1 EXEC PGM=&PGM

In JCL:

//MYJOB JOB (ACCT)

//PROC1 PROC PGM=IEFBR14

In this instance, I can override &PGM when calling, which comes in handy when reusing procedures across different teams.

36. What is the use of the NOTIFY parameter?

When I submit jobs, I often want to know the completion status of the jobs. The NOTIFY parameter sends completion messages to my TSO ID. Example:

//MYJOB JOB (ACCT),'NAME',NOTIFY=&SYSUID

This type of automatic notification system works to my advantage, which is especially appreciated on jobs that take a considerable amount of time to complete.

37. What are generation data group limits, and how do I manage roll-off?

By default, a GDG base can contain 255 generations. When it reaches its capacity, it simply rolls off the oldest generation. That is, let’s say if limit=5, and I create (+1), it creates a 6th generation but rolls off the 1st generation. I do this by setting the limit based on the business needs.

38. What is DISP=SHR vs DISP=OLD?

Here is the difference between DISP=SHR and DISP=OLD: 

• SHR: Allows concurrent read access. I use this when multiple jobs require the same input dataset.
• OLD: It provides exclusive control, meaning that until the job using the dataset file finishes, no other job can access it again. I use OLD when processing datasets to reduce the risk of someone else trying to use that dataset file at the same time.  

Making that distinction is very important for production jobs.

39. What are concatenated PROCLIBs in JCL?

A concatenated PROCLIB in JCL is a collection of Partitioned Datasets (PDSs) that are connected to create a single logical procedure library. This enables the system to look for procedures in a specified order across several libraries.

We have multiple PROCLIB DSNs in some projects. I concatenate them like this:

//JOB1 JOB (ACCT)

// JCLLIB ORDER=(PROCLIB1,PROCLIB2)

In this example, the system will search PROCLIB1 first, then PROCLIB2 for catalogued procedures. This offers some structure regarding which team or environment is associated with which documented procedures.

40. How do I use the TYPRUN=SCAN option?

To use the TYPRUN=SCAN option, you can place it in your job card. The TYPRUN=SCAN option will allow you to syntax check the JCL of the job without actually running the job. It’s handy for checking JCL for things like misspelt keywords, illegal characters, or unmatched parentheses before you submit the job for execution.

To use TYPRUN=SCAN, find your JOB statement. The first statement in your JCL is the JOB statement. Then type TYPRUN=SCAN. Add it as a keyword parameter as part of your JOB statement.

41. How do I override a dataset in PROC from the main JCL?

I override datasets by coding a DD statement in the calling JCL. For example, if the PROC has:

//STEP1 DD DSN=OLD.DATA,DISP=SHR

I can override in JCL like this:

//STEP1.DD DSN=NEW.DATA,DISP=SHR

I make this change to the JCL and do not modify the PROC so that I can use that PROC again in the future.

42. How do I use condition checking with abend codes?

Condition checking with ABEND codes, especially in a z/OS environment, enables us to execute JCL steps conditionally on whether or not a previous step has terminated abnormally. This is essential for directing job flow and executing recovery actions.

Methods of Condition Checking with ABEND Codes:

Checking Conditions with IF ABEND THEN Statements in JCL: 

• This is the contemporary “best practice” approach to conditional processing in JCL. // IF ABEND THEN 
• This checks if a previous step in the job stream terminated abnormally. If stepname.ABEND THEN 

This checks whether the stepname is terminated abnormally. 

• // IF (stepname.RC > 0 | stepname.ABEND) THEN 

This checks for either a return code of greater than zero or an ABEND in stepname.

Example: 

    //STEP1 EXEC PGM=MYPROG1

    // IF STEP1.ABEND THEN

    //STEP2 EXEC PGM=ABEND_RECOVERY_PROG

    // ELSE

    //STEP3 EXEC PGM=NEXTPROG

    // ENDIF

43. How do I optimise JCL jobs for better performance?

I optimise JCL jobs by:

• Using symbolic parameters: Symbolic parameters are variables that can be defined and used in JCL statements. They offer a method for changing the values of parameters rather than changing the code itself.
• Using conditional processing: Conditional processing is a process of controlling and optimising the JCL job, depending on the result of the previous step.
• Using restart and checkpoint features: Restart and checkpoint features are mechanisms that allow you to continue from a specific point of the JCL job if it is interrupted.
• Using optimal allocation and disposition: Allocation and disposition are parameters that define how the datasets used by the JCL job are created, accessed, and disposed of.
• Using performance tuning tools: Performance tuning tools are utilities that analyse and tune the performance of your JCL jobs.

//STEP1 EXEC PGM=IEFBR14

//NEWFILE DD DSN=MY.NEW.DATASET,DISP=(NEW,CATLG,DELETE),

//        UNIT=SYSDA,SPACE=(TRK,(5,2)),DCB=(LRECL=80,RECFM=FB,BLKSIZE=800)

//STEP2 EXEC PGM=XYZ,COND=(0,NE,STEP1)

// IF (STEP1.RC = 0) THEN

//STEP2 EXEC PGM=ABC

// ELSE

//STEP3 EXEC PGM=XYZ

// ENDIF

  //STEPNAME EXEC PROCNAME,PARM1=NEWVALUE1,PARM2=NEWVALUE2

// INCLUDE MEMBER=MYLIB(MYSTEP)

//INPUT DD DSN=DATA.FILE1,DISP=SHR

//      DD DSN=DATA.FILE2,DISP=SHR

//SYSUT1 DD DSN=&&TEMP,DISP=(NEW,DELETE),UNIT=SYSDA,SPACE=(CYL,(1,1))

//SYSPRINT DD SYSOUT=*,OUTDISP=PURGE