INTRODUCTION TO RECOVERY AND RESTART - IBM Mainframe

Online systems provide users with immediate access to the information they need to do their work. It is important that as a DB/DC control system CICS provide consistent and reliable access to that data. In order to meet this goal online system development must provide for the handling of errors that would result from power failure, system crash, or program abends, etc. So it is crucial to recover the resources and to restart the transactions or the entire system as quickly as possible. The goal of recovery processing is to:

  1. Maintain the integrity of the data processed by the system.
  2. Minimize the impact of a task or system failure on the system users.

The issue of recovery must be addressed throughout the system development life cycle, beginning in the early stages of design. Since recovery processing may have an effect on the performance of a system, it should be planned carefully to minimize the impact on system users.

CICS continually records information about the state of the region and about the state of each unit of work in the region. This information is preserved and used when a region is restarted, thus enabling CICS to restart with no loss of data integrity.

If CICS shuts down leaving any units of work in-flight, these units of work are backed out when CICS restarts and locks held by these units of work are released.

A CICS region can shut down in several ways:

Normal (or warm) shutdown

The CICS region shuts down in response to a CICS system management command. All transactions that have already started continue to run, and CICS waits indefinitely until they are complete. Long-running transaction can cause an unacceptable delay, or require intervention by the operator.

Immediate shutdown

The CICS region shuts down in response to a CICS system management command. All transactions are purged from the system; units of work are not backed out, and locks are not released, until the region restarts.

Abnormal shutdown

The CICS region shuts down in response to an unrecoverable error in a CICS system module, in response to a request from the operating system (arising, for example, from a program check or system abend), or following a machine check or power failure. In an abnormal shutdown, CICS might not be able to perform any shutdown processing. Units of work are not backed out and locks are not released until the region restarts.

Depending on whether the region has been started and shut down in the past, and on the way it was shut down, a CICS region can start in several ways:

Initial start

The CICS region starts with no reference to any previous activity. You perform an initial start when starting a new CICS region for the first time.

Cold start

The CICS region starts and restores recovery information that relates to systems and resource managers that were connected when the regions shut down. It does not recover the state of local resources.

Warm restart

A warm restart follows a normal shutdown and restores the region to its state it was in when shutdown completed.

Emergency restart

An emergency restart follows an immediate or abnormal shutdown of the CICS region. During an emergency restart, CICS backs out any transactions that were in-flight at the time of shutdown and thus frees any locks that were held at the protecting resources.

In normal operation, when a CICS region starts, it examines the information recorded before the region last shut down to determine in which way to start.


All rights reserved © 2018 Wisdom IT Services India Pvt. Ltd DMCA.com Protection Status

IBM Mainframe Topics