Browse Prior Art Database

Restore Write State for Check-Sum Direct Access Storage Device Subsystems

IP.com Disclosure Number: IPCOM000119306D
Original Publication Date: 1991-Jan-01
Included in the Prior Art Database: 2005-Apr-01
Document File: 3 page(s) / 96K

Publishing Venue

IBM

Related People

Styczinski, DA: AUTHOR

Abstract

A state of a check-sum storage subsystem, the restore write state, in which data can be restored more rapidly than in the normal protected state is disclosed. The act of returning to the protected state regenerates the check-sum. The total time to restore data, including the check-sum regeneration may be less than the time to restore directly while in the protected state.

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

Restore Write State for Check-Sum Direct Access Storage Device Subsystems

      A state of a check-sum storage subsystem, the restore
write state, in which data can be restored more rapidly than in the
normal protected state is disclosed.  The act of returning to the
protected state regenerates the check-sum. The total time to restore
data, including the check-sum regeneration may be less than the time
to restore directly while in the protected state.

      When data is being restored to a DASD subsystem utilizing
outboard check-sums (from a save file which may reside on tape or in
a separate direct access storage device (DASD) auxiliary storage pool
(ASP)), the operation may be completed more quickly by using a
special restore write state of the DASD subsystem rather than doing
check-sums as normal.  The operation of the restore write state is
the subject of this article.

      Use of the restore write state is appropriate when the system
is dedicated to the restore operation and normal user processing will
not resume until the restore is complete. The system puts the DASD
subsystem to be restored into the restore write state at the
beginning of the restore operation and puts the DASD subsystem back
into the normal protected state after the restore is complete.  While
in the restore write state, the DASD subsystem foregoes the check-sum
update and restore can thus proceed at the rate of unprotected
subsystems.  The process of returning the DASD subsystem from the
restore write state to the normal protected state causes the
check-sums to be updated.  The process of updating the check-sums
proceeds at the rate of sequential access of data on the DASD, which
is much faster than random access of the DASD.  Updating all the
blocks on the check-sum DASD takes 4 to 12 minutes depending on the
DASD characteristics.  All check-sum DASD can be updated at once so
that protection to the entire DASD subsystem can be restored in the
same 4 to 12 minutes.  For large restores to large subsystems, the
savings from using the restore write state followed by a 4 to 12
minute check-sum update will be substantial.

      DASD subsystems with outboard check-sums usually have a fast
write feature to improve performance.  The description of the restore
write state which follows incorporates use of the fast write buffer
(a nonvolatile RAM).  If...