Browse Prior Art Database

Nondisruptive Resuming of Work from a Checkstopped Central Processor

IP.com Disclosure Number: IPCOM000107761D
Original Publication Date: 1992-Mar-01
Included in the Prior Art Database: 2005-Mar-22
Document File: 2 page(s) / 88K

Publishing Venue

IBM

Related People

Daly, JC: AUTHOR [+2]

Abstract

Disclosed is a method whereby the MVS control program can immediately resume the work that had been running on a central processor (CP) that had checkstopped in a multi-CP system. BACKGROUND

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Nondisruptive Resuming of Work from a Checkstopped Central Processor

       Disclosed is a method whereby the MVS control program can
immediately resume the work that had been running on a central
processor (CP) that had checkstopped in a multi-CP system.
BACKGROUND

      In the past, CP checkstops always resulted in the MVS Alternate
CPU Recovery (ACR) function taking the CP  offline and initiating an
abend (OF3) against the unit of work that had been running on the CP.
The removal of a checkstopped CP is obviously appropriate but
questions arose as to why the abending of the work that had been
running on that CP was always necessary.

      The answer to this is simple - to resume, on another CP, the
work that had been running on a checkstopped CP, several thing are
required:
      1.  The hardware must be able to guarantee that it has
preserved the conditions and status for the work that was interrupted
by the checkstop to the point that it was before the inter
rupted instruction was started.
      2.  This status data from the checkstopped CP and an indicator
of the above "preservation" must be accessable to MVS running on the
CPs.
      3.  MVS must determine if the work on the checkstopped CP can
be resumed on another CP.
      4.  MVS must manipulate the status data provided by the
hardware to either immediately resume the work from the checkstopped
CP on another CP or to put it into some sort of queue for later
resumption.

      Since the hardware previously did not provide items 1 and 2,
MVS had no way to reconstruct the environment required to resume the
interrupted work on another CP so abending that work was the only
option.
PROBLEM SOLVED

      Since some types of machines can now provide the data in items
1 and 2 above, this article describes a new method by which MVS
identi fies whether the work on the checkstopped CP can be resumed on
another CP and, if so, how that work is moved to that other CP
without disruption.

      The objective of this process is to minimize the number of
abnormal terminations to units of work that a...