Job Control Language (JCL)

JCL

Is used to submit jobs to the system. Job is one unit of work. JCL consists of statements and their parameters. There are three types of statements: JOB (specifies the information of the job such as jobname, programmer name and programmerid), EXEC (specifies the program/procedure to use), DD (specifies the files to use in the job) and 2 types of parameters: positional (should be coded in specified order and position eg: accounting info should be coded after job keyword if not JCL error will come), keyword (can appear in any order)

JCL is useful to run the Batch processes.

JCL syntax

the record length should be 80 to write a JCL code: lrecl : 80, should be in upper case (caps on),

first 2 characters on every line is an identifier //,

name of the job: 8 characters it should be same as the name of the pds member,

operand: JOB/EXEC/DD

Accounting info:

other parameters like programmer name

//NAMEFIELD OPERAND PARAMETERS

comment: //* writing comment in the same line /*comment

end jcl : /*

Job Statement

Also known as job card.

line identifier //

name of the job: sortjob - 8characters max, can be alphanumeric, first character should never be a number

JOB operand

accounting info: refers to the person or group to which the CPU time is owned. System specific. not mandatory. used to uniquely identify a user can be emp id or name of dept. multiple (1234, ‘HR’). To skip this parameter we can give ,

programmer name: not mandatory

keyword parameter - can come in any order:

Class= can be A to Z or 0 to 9. every class defines a certain function. jobs similar in some aspect are grouped together in one class For eg: class P is for print. class=Z, class=1

priority= PRTY: 0 to 15. 15 is the highest and 1 is the least.

notify= which person to notify when job completes? - our ID. can also give &sysuid (person who submits the job will get notified)

message class: MSGCLASS= a to z or 0 to 9. specifies the o/p destination for the job messages. P - print the output. A- could mean fax the job output, D - could mean email the output

message level: MSGLEVEL=(Statement, Msg) Specifies the type of messages to be written to the o/p. message level can be 0,1,2, message can be 0, 1

Statement- 0, job statements only

1, JCL along with symbolic parameters expanded

2, JCL only

MSG- 0, messages written only on abnormal termination

1, messages written irrespective or normal/abnormal termination

Type run: TYPRUN: scan(only scan the JCL for syntax errors and don’t run) or hold(hold the job don’t execute it can be executed if required )

Time: max time the job can execute TIME=(50,00) 50 minutes

region: specifies the amount of address space that will be accolated to the job execution. nK(kilobyte) or nM(megabyte) region=5K

When a parameter is omitted it takes up the default system value which was defined by the system administrator.

exec statement

Used to invoke the pgm/proc that we want to execute

Tells system what pgm/proc to execute.

One JCl can have multiple EXEC statements

syntax: //stepname EXEC parameters

stepname - 1 to 8 characters, coded from column 3. First character must be alphabetic or nation character ($ , # , @)

positional parameters used:
PGM - it refers to the program name to be executed in the job step

PROC - it refers tot he procedure name to be executed in the job step

a procedure refers to a set of JCL statements that are stored in a library and retrieved by its name.

a procedure may contain one or more EXEC statements.

Keyword parameters:

PARM: used to pass values to the pgm being executed

ACCT: specifies the accounting info of the job step. ACCT=(userid)

TIME: MAximum time allocated for the step execution. TIME=(mm,ss) of TIME=ss

region: Specifies the amount of address space to be allocated for the job execution. REGION = nK | nM

When parameters are omitted default is used

DD statements

DD - Dataset definition

Define all the datasets to be used in the JCL.

Syntax: //DDname DD parameter

DDname must be 1-8 characters long, coded from column 3, first char must be alphabetic or national char(@,#,$)

Positional Parameters:

‘*’ : * begins an in-stream data set

Dummy: used for testing pf a JCL. Tells the system to not use any i/p/o/p for this DD statement

Data: In-stream dataset containing statements, which begin with //

Keyword parameters:

DSN: Dataset Name (DSN) is used to give the name of the Dataset. DSN= PHYSICAL Dataset name

DISP: Disposition describes the status of the dataset, what to do with the dataset on normal end of the job and what to do with the dataset on the abnormal end of the job.

DISP=(status, normal end disposition, abnormal end - disposition)

status: NEW- dataset will be created by the DD statement

OLD- dataset already exists, existing data will be overwritten and the job step will have exclusive access until completion of job set

SHR - dataset already exists and will be read, multiple jobs can access it at the same time

MOD - dataset already exists, if not then new dataset will be created. Data will be added at the bottom of the existing data.

normal end - CATLG, UNCATLG, DELETE, KEEP, PASS

abnormal end - CATLG, UNCATLG, DELETE, KEEP

CATLG- dataset is retained with an entry in the system catalog

UNCATLG- dataset is retained but the system catalog entry is removed.

KEEP- Dataset is retained without changing any of the catalog entries

DELETE- dataset is deleted from user and system catalog.

PASS- used for normal end only. this is used when the dataset is to be passed and processed by the next job step in a JCL

Default value: DISP=(NEW,DELETE,DELETE)

DCB: Data control Block (DCB) specifies the values for creating a new dataset

It contains various parameters:

LRECL: record length of the dataset

RECFM: record format of the dataset such as FB, U,VB

BLKSIZE: Blocksize of the dataset

DSORG: Dataset organization (PS(for PS), PO (for pds))

Space: Specifies the space required for the dataset.

SPACE=(spaceunits, (primary, secondary, directory-blocks),RLSE)

UNIT: specifies what type of storage to use for the dataset

UNIT=DASD | SYSDA

DASD will store the dataset in Direct access storage device which is like a tape storage

SYSDA will store it in next available disk storage type

VOL: specifies the volume number on the device where the dataset needs to be stored. VOL=SER=(v1,v2). vi,v2 are volume serial numbers

SYSOUT: specifies the o/p destination based on the class. SYSOUT = class, where class is A then o/p is printer. * then o/p goes to the same destination as that of the MSGCLASS.

JOB processing

JCL Submission → Job entry subsystem (JES) → JCL interpretation in JES ( JCL is converted into text that JES understands and stored in dataset called as spool) → JES decides the priority based on class and prty. If no JCL errors then job is queued → Job execution and log creation in spool.

JES

Job Entry Subsytem is used to:
1) submit jobs to the system

2) schedule the processing of the job

3) Control the output of the jobs

SPOOL

Simultaneous Peripheral Operations Online is a dataset which stores the:

1. Interpreted text for the JCL execution

2. messages related to job execution, such as job log

3. output data from the job execution such as errors

system display search facility (SDSF) is used to view spool.

Return codes

When a job executes a return code is set based on the status of execution. Cav vary from 0 to 4095.

0 - Successful execution

4 - Successful execution with warnings

8 - Errors

12 - Severe Error

16 - Major Error

MAXCC is the maximum of the return code from all of the job steps.

SDSF

System Display and Search Facility is used to review the job output. Its primary purpose is to display printed output in the spool area.

SDSF can be used to:

1. View and search the system log

2. hold, release, cancel and purge jobs

3. monitor jobs while they are processed

4. display job output

5. control the order in which jobs are processed

SDSF can be used to save the data to a dataset for backup.

SDSF commands

TSO SDSF to go to spool or use 13.14.

ST to see all jobs

? - List the datasets

s jobname - search jobname

Owner userid - show jobs owned by userid

owner * - show all the jobs

f jobname - find and locate to jobname

pre jobname - prefix of jobname eg - pre sor* - will list all the jobs starting with sor

P - purge the dataset - that job log in spool will be deleted

C - cancel the job

xdc - store the output to a dataset

JESMSGLG - message log, execution time, when was it submitted, user who submitted it, class executed- basic info on the job

JESJCL - will have the whole JCL

JESYSMSG - system message - has all details step by step

SYSOUT will have info on each step separately. SYSOUT step01, SYSOUT step02 etc.

Instream Procedure

Procedure (its like function, way to reduce the lines of code by writing the redundant code in procedure) which is defined in the same JCL.

SYntax:

//procname PROC

//steps

// PEND

calling:

//step2 exec procname

Symbolic Parameters

Used to pass values to the procedure. Always starts with an &. We can give any name except reserved keywords.

//procname PROC

//steps EXEC PGM=IEFBR14

// DD1 DD DSN=&DNAME,

// DISP and other parameters

// PEND

calling:

//step2 exec procname,DNAME=TRAIN04.TEST1

//step3 exec procname,DNAME=TRAIN04.TEST2

can be used to pass value for part of the name too- like just one qualifier:

//DD1 DD DSN=TRAIN04.&MNAME..PS1

calling:

//step2 exec procname,MNAME=TEST1

starts with & and ends with . or we can also use space: TRAIN04.&MNAME..PS1

if its the last qualifier then no need for . TRAIN04.&MNAME

Cataloged Procedures

Cataloged procedure will be in another location and we can call it in JCL

The location has to be mentioned in the jcl. JCLLIB is the kwyword. that proclib is the stepname can be anything

//PROCLIB JCLLIB ORDER=TRAIN.TEST.JCLLIB

//STEP1 EXEC CREPSPR,DNAME=TRAIN04.TEST.PS1

Nested Procedure

Calling another procedure within a procedure. We can have upto 15 levels of nesting:

eg:

//CREPSPR PROC

//STEP1 EXEC PGM=IEFBR14

//DD…

//STEP2 EXEC ANOPRC,FNAME=TAY /* let this is another proc

// PEND

JOBLIB - JOB LIBRARY

Its for load module, its executable. Default location of load modules is SYS1.LINKLIB. When we don’t specify the load lib, JCL will first check the default location, if not found it will give abend S806.

to solve it we have to specify the load module. JOBLIB is used and it is given after the JOB CARD and is applicable to whole JCL

//JOBLIB dd dsn=TRAIN04.TEST.LOAD,DISP=SHR

So now it will search the PGM in the specified location and if its not found then it will search in default.

multiple dataset loc can be given:

//JOBLIB dd dsn=TRAIN04.TEST.LOAD,DISP=SHR

// dd dsn=TRAIN04.TEST.LOAD1,DISP=SHR

// dd dsn=TRAIN04.TEST.LOAD2,DISP=SHR

STEPLIB

Its for load module, its executable. STEPLIB is given after the step. It will find the location of the program specifeid in the steplib. If JCl have both JOBLIB and STEPLIB then steplib will be used and joblib will be ignored. It is applicable to that step only

//STEP1 EXEC PGM=PGMA

//STEPLIB DD DSN=TRAIN.CRS.LOAD,DISP=SHR

// dd dsn=TRAIN04.CRS.LOAD1,DISP=SHR

JCLLIB

Its for JCL code. Used to tell the location of JCL code

//stepname JCLLIB ORDER=TRAIN04.TEST.JCLLIB

we can give more than 1 loc:

//proclib JCLLIB ORDER=(TRAIN04.TEST.JCLLIB,TRAIN04.TEST2.JCLLIB)

COND parameter

It is used to decide whether to execute a step or not.

Cond= (returncode, operator, stepname)

eg: COND=(0,EQ,STEP01) - if the cond is true then step is bypassed and if it is false then it executed.

COND=(8,LE)

operators: GT, LT, EQ, NE, GE (greater than or equal), LE (Less than or equal).

COND can be given in a JOB statement or an EXEC statement. COND code does not work on the first step in a JCL.

COND can be given in the job statement if that cond has to be applied to every steps. //NATARKIJ JOB 1,NOTIFY=&SYSUID,COND=(0,EQ) - it will be applied all the steps except the first step.

Never execute a step/ to skip a step - COND

COND=(0,LE) - this step will always be true thats why the step will not execute

to always execute a step - COND

COND=(4095,LT) - this step will always be false thats why the step will execute

COND=ONLY

the step will only execute if any of the previous steps has abended.

COND=EVEN

execute even if abend occurs.

IF..ELSE in JCL

//check1 if rc = 0 then

//step2 exec pgm=sort

// endif

if else:

//check1 if rc = 0 then

//step2 exec pgm=sort

//else1 else

//step3 exec pgm=sort

// endif

for abend

//check1 if abend then

if else will work on first step of JCL and COND parameter doesn’t.

GDG

Generation Data Group is a group of files related to each other and having a common name.

eg: a bank can have datasets containing the names of its new customers in JAN,FEB and so on. There will be twelve datasets per year for the bank. These datasets will be grouped together to form a GDG instead of twelve different datasets

BANK.CUST.MON.G0001V00

BANK.CUST.MON.G0002V00

G001 stands for generation number

V00 stands for version number - version for the same version

GDG base→ to create a GDG, a dataset called GDG base is first created. All the generations are then modeled from this base.

BANK.CUST.MON → GDG base

Max np. GDG generations - 255

to access the latest generation - BANK.CUST.MON(0)

second most recent - BANK.CUST.MON(-1) and the rest so on

IDCAMS is used to define/alter/delete a GDG.

Parameters:

Limit - max number of generations that a GDG can hold.

EMPTY - uncatlogs all generation once limit is reached.

NOEMPTY - uncatolgs only the oldest generations once limit is reached

SCRATCH - deletes the generation when uncataloged

NOSCRATCH - does not delete the generation when uncatologed.

Create a GDG

//STEP01 EXEC PGM=IDCAMS

//SYSPRINT DD SYSOUT=*

//SYSIN DD *

DEFINE GDG(NAME(TRAIN04.TEST.GDGGEN) -

LIMIT(12) -

NOEMPTY -

SCRATCH

to create generations:

//STEP01 EXEC PGM=IEFBR14

//DD1 DD DSN=TRAIN04.TEST.GDGGEN(+1),

// DISP=(NEW,CATLG,DELETE),

// SPACE=(TRK,(100,100),RLSE),

// DCB=(LRECL=80,RECFM=FB,BLKSIZE=800,DSORG=PS)

//SYSPRINT DD SYSOUT=*

to delete:

//DD1 DD DSN=TRAIN04.TEST.GDGGEN(0),

// DISP=(OLD,DELETE,DELETE) referencing the gdg:

to reference a GDG gen in the same JCL it is created then we have to give TRAIN04.TEST.GDGGEN(+1), disp=shr

latest version not created in the same JCL- TRAIN04.TEST.GDGGEN(0), disp=shr

second most- TRAIN04.TEST.GDGGEN(-1), disp=shr and so on

Delete GDG

//STEP01 EXEC PGM=IDCAMS

//SYSPRINT DD SYSOUT=*

//SYSIN DD *

DELETE GDG(NAME(TRAIN04.TEST.GDGGEN) GDG FORCE

/*

we can also give PURGE in place of FORCE. PURGE will delete even if the retention period of the GDG is not expired. Force will delete will only delete if retention period has expired.

REXX

/* REXX */

Say ‘HELLO WORLD’

say ‘Enter you name: ‘

PULL name

a = 0

DO until a = 5

say ‘this is’ A ‘times printing’

a = a+1

end

execute - 1) give → ex in front of dataset

2) go to 6 in ispf → ex ‘rexx.code(hello)’

3) tso ex ‘rexx.code(hello)’

4) go to ready prompt and give → ex ‘rexx.code(hello)’

VSAM

Virtual Storage Access Method. VSAM is a data access method in Mainframe.

Access method: it defines the technique that is used to store and retrieve data

An access method has:

1) its own structure to organize data - in VSAM→ Data and index component. Data is organized in control intervals

2) System-provided programs (or macros) to define datasets - IDCAMS

3) utility program to process datasets - IDCAMS

IDCAMS is used to define VSAM datasets

VSAM can’t be viewed using ISPF

VSAM are also called clusters.

Control intervals- VSAM stores and retrieves data from disk as control intervals. In every read one CI is read from the disk and put into the memory

CI has record record record record free space RDF RDF CIDF

Records, free space (free space is the % of free space reserved in a CI and CA), CI fields: RDF(Record Definition Field) is 3 bytes long, it describes the length of records and CIDF (Control Interval Definition Field) is 4 bytes long, it contains the info about the CI

Default size of CI is 4096 bytes. Max of 32kb.

Control Area(CA): two or more CY are grouped together to form a CA.

VSAM size is always a multiple of CA:

eg: if CA = 8192 KB then VSAM = n*8192 kb

index consists of sequence set and index set. every control area has a seq set associated with it. every record in seq set points to one CI. Index set points to seq sets. seq set useful for sequential processing. Every record of seq set stores the last address of the CI. every record in index set will point to last address of seq set. Index set is used for direct access and seq set is used for sequential access

Free space= %CI to be free, %CA to be free → eg: freespace =(20,25) means 20% of every CI and 25 % of the whole CA will be free. Its useful to insert data in VSAM. It is provided during the creation of the VSAM

VSAM splits: when a record is being inserted into a VSAM which has no more space in the CI, then a Ci split occurs. done the area where record has to be inserted. It creates free space.

QSAM

Access method on Mainframe. PS files are created and stored using QSAM.

QSAM stores data as blocks. A block is a collection of records. In every read whole block is copied into the memory.

This equivalent to Control Interval in VSAM. Records are stored and retrieved from the disk as Control Interval (CI). In every read one CI is read from the disk and put into memory.

Types of VSAM

KSDS - every record is retrieved and inserted using a key value. Thus it has 2 components: Data and Index.

Data has actual Data inside whereas the index points to the location of the data.

RRDS - the record can be accessed using record number

ESDS - the records are retrieved in the sequence they were added. just like PS
LDS - the data is stored in byte-streams. Like notepad

KSDS

Key-Sequenced Data set- records are stored in ascending seq by key. Always sorted by the key. keys are unique. No duplicate records are possible in KSDS.

Parameters to create KSDS:

i) Name

2) keys: its syntax is keys(length offset) for eg: if key is 23413 - length = 5 and offset is 0 i.e. the index position which we want to be key. so keys will be keys(5 0)

suppose we want 34 in 23413 be the key then keys will be keys(2,1). max length of key 255.

3) RECSZ: length of most of the records. record size→ RECSZ(average maximum)

if a record has fields emp no, emp name and address and approx. length of each field is 5 + 30 +53 then avg is 90 and max maybe 200.

4) FREESPACE (%CI %CA)

5) INDEXED: it tells the system that the VSAM is KSDS. It is NUMBERED for RSDS, NONINDEXED for ESDS and LINEAR for LDS

6) CISZ (Control Interval Size): Default is 4096 Bytes. Can be a maximum of 32K

7) SPACE: The amount of total space allocated to VSAM. Syntax: Unit(primary secondary) eg: Tracks(30 50)

8) VOLUME: The volume serial where VSAM will reside is given here.

create KSDS JCL

//STEP01 EXEC PGM=IDCAMS

//SYSPRINT DD SYSOUT=*

//SYSIN DD *

DEFINE CLUSTER( -

NAME(TRAIN04.TEST.KSDS) -

INDEXED -

KEYS(5 0) -

RECSZ(90 200) -

FREESPACE(10 20) -

TRACKS(50 30) -

CISZ(8192) -

volume(ZASYS1))

/*

ESDS

Entry-sequenced Data set, records are stored in the order they are added