IMS Startup

This section describes the types of IMS system starts that can be performed for an IMS system and how to start IMS-associated regions.

Types of IMS System Start

This section describes the more common types of IMS system starts that can be performed. These types of IMS system starts are applicable to both IMS TM and IMS DB.

Cold start

An IMS control region cold start is done the first time you start the system. During a cold start, IMS initializes the message queues, the dynamic log, and the restart data sets.

Normal restart

Normal restart (or warm start) is the restart of IMS after a normal IMS termination. A normal restart preserves the message queues.

Emergency restart

If IMS terminates abnormally, IMS performs an emergency restart with the logs that were active at the time of failure. Emergency restart processing backs out the full-function database changes of incomplete transactions. The output messages inserted into the message queues by these incomplete transactions are deleted.

After back out processing finishes, the input messages are enqueued and the pending output messages are transmitted again. Application programs must be restarted manually. If a BMP or JBP application was active at the time of failure, it must be resubmitted by using z/OS job management. If the JBP or BMP application uses extended restart calls, which are used to restart the application from a checkpoint and not from the beginning, it must be restarted from its last successful checkpoint to avoid missing or inconsistent output.

Automatic restart

With an automatic restart, IMS starts up using either a normal restart or an emergency restart, depending on the status of the previous shutdown.

If the last IMS shutdown was successful, a normal restart will be performed. If the last IMS shutdown was abnormal (from a failure), IMS automatically performs an emergency restart.

For most installations, automatic restart should be the default (specify AUTO=Y in the control region JCL).

Other types of manual restarts

Numerous other types of manual restarts are possible with IMS, each with unique requirements. For detailed information about these other types of restarts, see IMS Version 9: Operations Guide and the IMS Version 9: Command Reference.

Starting Regions That Are Related to IMS

The following sections describe how the various IMS-related regions are started.

Address Spaces

All address spaces can run either as a started task or as a job. In most cases, the IMS control region and the separate address spaces will run as started tasks. The application dependent regions are run as either jobs or started tasks.

When a control region is started, it issues a z/OS START command to start the DLISAS and DBRC regions, as shown in the following example:

 START xxxxxxxx, PARM=(DLS,imsid) START xxxxxxxx, PARM=(DRC,imsid) 

The xxxxxxxx fields are the procedure names. These commands will start the DLISAS and DBRC regions, respectively.

Starting Application Dependent Regions

IMS will not automatically start application dependent regions. There are several ways to start these regions.

• The IMS time controlled operations (TCO) can issue /START REGION commands. TCO is a time-initiated IMS facility that can generate any valid operator IMS input.

• Some forms of automation programs can issue either IMS or z/OS START commands.

• A job scheduling system can submit jobs based on time or the notification of IMS being started. The notification can be in the form of automated messages.

Message Processing Regions

IMS MPR regions are normally started by an IMS start region command, as shown in the following example:

 /START REGION xxxxxxxx 

The xxxxxxxx field is the member name in a library. The members contain the jobs for the MPR regions. The IMSRDR procedure is used if the MPRs are jobs. The IMSRDR procedure is customized to point to the correct library to find the job JCL. If you run multiple IMS subsystems on a single z/OS system, they normally use a different version of the IMSRDR procedure, with each pointing to different libraries. The procedure name is specified on the IMSCTF macro in the system definition.

Related Reading: For details about the IMSRDR procedure or the IMSCTF macro, see IMS Version 9: Installation Volume 2: System Definition and Tailoring.

IMS Fast Path Application Regions

IMS Fast Path (IFP) regions are normally started in a similar fashion as MPR regions and follow the same rules and procedures.

Batch Message Processing Regions

Batch message processing (BMP) regions are almost always started outside of IMS. Most BMP regions are scheduled at appropriate times to meet application requirements. As long as the IMS control region is available, the BMP regions can run. BMP regions can execute even though there are no MPRs running at the time.

Java Non-Message Driven Application Processing Region

This region, which is called a Java batch processing (JBP) region, is similar to a BMP region. The JBP region is started in the same manner as a BMP region. The default job name is IMSJBP.

Java Message-Driven-Application Processing Region

This region, which is called a Java message processing (JMP) region, is similar to an MPR region. The JMP region is started in the same manner as an MPR region. The default job name is IMSJMP.