Transaction Lifecycle

A CICS transaction is born when a task starts, lives while the program runs, and dies when the program returns control for the last time. The diagram below shows the full lifecycle.

How a transaction starts

• The user types a trans-id, or a program issues a START or XCTL command.
• CICS creates a task — the runtime instance that will run the program.
• CICS loads the first program named on the TRANSACTION definition and gives it control.
• On the very first run there is no COMMAREA yet, so EIBCALEN is zero and the program shows the first screen.

What happens while it runs

• The program talks to the terminal with SEND MAP and RECEIVE MAP, and reads or updates files and databases.
• Every EXEC CICS command is a service request that CICS handles for the task.
• In a pseudo-conversational design the program does its work for one screen, then ends the task with RETURN TRANSID.
• The task ends but the transaction is not over — CICS remembers which trans-id to restart when the user responds.

How a transaction ends

• Pseudo-conversational step: RETURN with the TRANSID option ends the task and frees all its resources; the terminal is free for other users while this user thinks.
• When the user presses ENTER or a PF key, CICS starts a new task for the same transaction and restores the COMMAREA.
• Final end: RETURN without the TRANSID option ends the transaction completely — no restart happens.
• A conversational program instead just waits for the user with the task alive, which wastes CICS resources and is avoided in good design.
• Example of the RETURN that continues the lifecycle:-
  
  EXEC CICS RETURN TRANSID('MNT1') COMMAREA(CA-AREA) END-EXEC. *--- Task ends here. The next keypress restarts the * transaction as a brand-new task with CA-AREA restored ---*