This topic describes the
three types of CICS® system shutdown (normal, immediate, or uncontrolled)
and the events that cause them. For information about CICS shutdown, see the CICS Recovery and Restart Guide.
- In normal shutdown, CICS performs a controlled sequence of operations
that leave the system in a well-defined state. Existing tasks are allowed
to finish.
The following events can cause normal shutdown of CICS:
- Using the CEMT PERFORM SHUTDOWN transaction
- Using the EXEC CICS PERFORM SHUTDOWN command.
Note:
During shutdown the trace subtask TCB will be
detached. This will cause ABEND13E. This is normal processing, and is to be
expected.

- In immediate shutdown, CICS remains in overall control, but it does a
minimum of processing so the system can be terminated rapidly. Existing tasks
are not allowed to finish, and could abend. If the CESD default shutdown transaction
is enabled, existing tasks are given a short time to finish before they are
purged.
The following events can cause immediate shutdown of CICS:
- Using the CEMT PERFORM SHUTDOWN IMMEDIATE transaction
- Using the EXEC CICS PERFORM SHUTDOWN IMMEDIATE command
- A CICS system abend
- A program check.
Note:
During shutdown the trace subtask
TCB will be detached. This will cause ABEND13E. This is normal processing,
and is to be expected.

- In uncontrolled shutdown, CICS is not given the chance to do any processing
after the event causing it to terminate has occurred.
The following events
can cause uncontrolled shutdown of CICS:
- Power failure
- Machine check
- Operating system failure.
After a normal shutdown, it is possible to warm start CICS. After an immediate
or an uncontrolled shutdown, an emergency restart or a cold start must be
performed.
When an XRF active CICS region is terminated abnormally, the
alternate CICS region normally completes initialization and takes over. However,
the alternate CICS region also terminates if the user has initiated a normal
shutdown of the active CICS region and has not specified
that takeover is to take place.
Normal shutdown is initiated
by the master terminal operator or by an application program, and is accomplished
in phases. (In comparison, immediate shutdown is accomplished by termination
processing.)
First stage of normal shutdown
During the first stage
of CICS normal shutdown, all terminals are active and all CICS facilities
are available. The following actions take place concurrently:
- Message DFHTM1715 is issued to the console and the master terminal user
to inform the operator that CICS is terminating.
- Tasks that already exist will complete. (Long running tasks, such as conversational
tasks, must end before this stage of shutdown can complete.)
- Tasks to be automatically initiated will run, if they can start before
the second stage.
- Any user-written programs listed in the first part of the shutdown program
list table (PLT) are run sequentially.
- The Front End Programming Interface (FEPI) is requested to shut down.
- The terminal that initiated the shutdown, if any, is detached. This allows
the operator to start any further tasks that might be required, or to purge
any tasks.
A new task
is allowed to start only if it has been defined as SHUTDOWN(ENABLED) in its
TRANSACTION resource definition, or, for a transaction started as a result
of terminal input, if the transaction identifier is listed in the current
transaction list table (XLT). The XLT list of transactions restricts the tasks
that can be started by terminals and allows the system to shut down in a controlled
manner. The current XLT is the one specified by the XLT=xx system initialization
parameter, which may be overridden by the XLT option of the CEMT or EXEC CICS
PERFORM SHUTDOWN command.
Certain CICS-supplied transactions are, however,
allowed to start whether or not their code is listed in the XLT. These transactions
are CEMT, CESF, CLS1, CLS2, CSAC, CSTE, and CSNE.
Note:
You
should not change the SHUTDOWN(ENABLED) attribute of the resource definitions
for these transactions, otherwise CICS may not shut down successfully.
- A request is issued to all interregion communication (IRC) activity.
- Terminal control is requested to ignore all further input.
- Unless SDTRAN=NO or NOSDTRAN was specified, the shutdown task starts the
specified shutdown transaction (default is CESD). CESD manages the purging
of long-running user tasks.
- If this is a non-XRF system, CLSDST requests are issued for all VTAM® terminals.
- The termination task waits for all terminal activity to cease, before
entering the second stage of shutdown.
The first shutdown stage is complete when the last of the programs specified
in the first part of the shutdown PLT has run and all user tasks are complete.
Second stage of normal shutdown
During the second stage of shutdown, terminals are not active,
and no new tasks are allowed to start. The following processing takes place:
- User-written programs listed in the second part of the shutdown PLT (if
any) are executed sequentially. These programs cannot communicate with terminals,
or make any request that would cause a new task to start.
- All currently open CICS files are now closed.
- The transient data CI buffer and the temporary storage buffers are flushed.
- CICS writes the following information to the global catalog:
- A warm keypoint. This contains information that is used to restore the
operating environment during a subsequent warm start.
- A warm-start-possible indicator. This status applies on the next initialization
of CICS if START=AUTO is specified.
- Transient data is terminated.
- A dump is taken, if one is required.
- If TAKEOVER was specified on the command to shut down an XRF CICS region,
a "signoff abnormal" request is made from the CICS availability manager
(CAVM).
- The local and global catalogs are closed.
- The following message is issued:
DFHKE1799 applid TERMINATION OF CICS IS COMPLETE
- CICS completes some internal processing, then returns control to MVS™.
During immediate shutdown
of CICS, possibly requested by the master terminal operator or an application
program, processing is different from a normal shutdown in the following important
ways:
- User tasks are not guaranteed to complete for any kind of shutdown. They
are just given less time for immediate shutdown before being purged.
- None of the programs listed in the shutdown PLT is run.
- CICS does not write a warm keypoint or a warm-start-possible
indicator to the global catalog.
- CICS does not close files defined to CICS file control.
To preserve data integrity, the next initialization of CICS must be an emergency restart. If the next initialization of CICS specifies
START=AUTO, there will be an emergency restart.
The processing involved in immediate shutdown is described as CICS system
termination processing. (In comparison, normal shutdown involves quiesce processing.)
Unlike processing, controls are not exercised to ensure that resources
and services remain available as long as they are needed. One consequence of this is that transaction
and CICS system abends can occur during immediate shutdown. Thus, if a task
tries to use a resource that has already been terminated, the task abends. Then dynamic
transaction backout is invoked, and that might also fail because it could
also try to use a resource that has been terminated.
In addition, if CICS system termination processing is delayed significantly,
tasks in the system waiting for input from terminals that are no longer available
are likely to extend beyond the period for deadlock timeout specified in the
DTIMOUT option of the TRANSACTION definition.
First stage of immediate shutdown
During the
first stage of an immediate shutdown, the following processes take place:
- The system termination task drives the collection of termination statistics.
- If there is a terminal associated with the event that caused the immediate
shutdown, a message is sent to inform the operator that CICS is terminating.
- If the shutdown request has arrived by transaction routing, the associated
terminal is freed.
- Terminal input is no longer accepted.
- The Front End Programming Interface (FEPI) is requested to shut down immediately.
Unless SDTRAN=NO or NOSDTRAN was specified, the shutdown task starts the specified
shutdown transaction (the default is CESD). CESD manages the purging of long-running
user tasks.
Second stage of immediate shutdown
During the
second stage of an immediate shutdown, the following processing takes place:
- Transient data is terminated.
- A dump is taken, if requested.
- Interregion sessions are terminated.
- If CICS is signed on to the CICS availability manager (CAVM), a "signoff
abnormal" request is made from CAVM.
- The local catalog and global catalog are left to be closed by the operating
system.
- The following message is issued:
DFHKE1799 applid TERMINATION OF CICS IS COMPLETE
- CICS completes some internal processing, then returns control to MVS.
An uncontrolled shutdown
of CICS can be caused by a power failure, a machine check, or an operating
system failure.
In each case, CICS cannot perform any shutdown processing. In particular,
CICS does not write a warm keypoint or a warm-start-possible
indicator to the global catalog.
To preserve data integrity, the next initialization of CICS must be an emergency restart. If the next initialization of CICS specifies
START=AUTO, there will be an emergency restart.
[[ Contents Previous Page | Next Page Index ]]