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

Automatic Reconfiguration of Disk Arrays

IP.com Disclosure Number: IPCOM000110478D
Original Publication Date: 1992-Nov-01
Included in the Prior Art Database: 2005-Mar-25
Document File: 1 page(s) / 60K

Publishing Venue

IBM

Related People

Mattson, R: AUTHOR [+5]

Abstract

Disclosed is a technique for automatically reconstructing data on a failed disk in a disk array. This technique, on a failure, combines two disk arrays in the system to build a larger disk array. This eliminates the need for providing spare disks in the system for reconstructing the data on a failed disk. This technique allows tolerating as many failures as the number of arrays in the system compared to the earlier techniques where the number of tolerated failures is limited to the number of spares provided in the system.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 64% of the total text.

Automatic Reconfiguration of Disk Arrays

      Disclosed is a technique for automatically reconstructing data
on a failed disk in a disk array.  This technique, on a failure,
combines two disk arrays in the system to build a larger disk array.
This eliminates the need for providing spare disks in the system for
reconstructing the data on a failed disk.  This technique allows
tolerating as many failures as the number of arrays in the system
compared to the earlier techniques where the number of tolerated
failures is limited to the number of spares provided in the system.

      In disk arrays, failure of a disk requires that the data on the
failed disk be copied to spare space to avoid losing data due to
another failure.  In earlier systems, spare disks or spare space is
provided to allow reconstructing the data on a failed disk.  The
technique proposed here uses the parity space of another disk to
reconstruct the data on the failed disk.

      Consider a system with two disk arrays, each with two data
disks and a parity disk, as shown in Fig. 1.  The basic idea behind
the technique is to merge the two arrays in the system into one such
that the combined array is still protected by parity.  The
reconfigured array would be as shown in Fig. 2.  By merging the two
arrays, all the data disks are still protected by parity and, hence,
the resulting system can tolerate another failure without losing any
data.

      Fig. 1 shows a system where parity is stor...