Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Refresh/Recovery Methodology for a Distributed Operation

IP.com Disclosure Number: IPCOM000103821D
Original Publication Date: 1993-Feb-01
Included in the Prior Art Database: 2005-Mar-18
Document File: 2 page(s) / 57K

Publishing Venue

IBM

Related People

Duvall, KE: AUTHOR [+5]

Abstract

Disclosed is a method to provide recovery for backup and restore processes that take hours to complete. In a distributed, multiple workstation, client/server based environment, backup and restore operations are time critical. If a failure occurs during a long duration operation (many files or a large volume of data is to be transferred), most technologies 'back-out' the partially completed operations and force the user to completely reinitiate the operation.

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

Refresh/Recovery Methodology for a Distributed Operation

      Disclosed is a method to provide recovery for backup and
restore processes that take hours to complete.  In a distributed,
multiple workstation, client/server based environment, backup and
restore operations are time critical.  If a failure occurs during a
long duration operation (many files or a large volume of data is to
be transferred), most technologies 'back-out' the partially completed
operations and force the user to completely reinitiate the operation.

      In this environment, a restart operation is defined.  This is
done by tracking the status of a particular request in two distinct
files:  the Status File and the Status Table.  The Status File
maintains a list of files that will be moved to the Archive Server
(backup data repository).  The Status Table maintains state
information about each request.

      Overall recovery is based upon state information provided by
each component of the product.  At the discrete states 'start' and
'end' entries are written to the Status Table to indicate the current
state of a request.  Depending upon the current state of the failed
request, the application will attempt to recover the failed request.
When a failure occurs, the restart function searches the Status Table
for a given primary request identifier, and based upon that search it
determines where the operation was prematurely terminated.  The
restart operation then initiates the completion of the original
request by reactivation of the original request identifier based on
data structures written to the Status File.

      In t...