Browse Prior Art Database

Virtual Hot Spare Disk Units

IP.com Disclosure Number: IPCOM000113515D
Original Publication Date: 1994-Sep-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 4 page(s) / 156K

Publishing Venue

IBM

Related People

Larson, WA: AUTHOR [+3]

Abstract

When a disk unit attached to a computer system (Fig. 1a) fails, the data stored on that disk unit may be lost and/or availability of the computer system to do useful work may be temporarily lost. If the disk unit is protected using mirroring (Fig. 1b) or RAID (Fig. 1c) techniques, the protection becomes exposed when a disk unit fails, and then data and/or system availability may be lost if a second disk unit fails prior to the repair or replacement of the first failed unit. The potential for data or system availability loss may be reduced if a spare disk unit is attached to the computer system. A method is described for providing the spare disk unit capability without actually requiring that a specifically designated spare disk unit be attached to the computer system.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 40% of the total text.

Virtual Hot Spare Disk Units

      When a disk unit attached to a computer system (Fig. 1a) fails,
the data stored on that disk unit may be lost and/or availability of
the computer system to do useful work may be temporarily lost.  If
the disk unit is protected using mirroring (Fig. 1b) or RAID (Fig.
1c) techniques, the protection becomes exposed when a disk unit
fails, and then data and/or system availability may be lost if a
second disk unit fails prior to the repair or replacement of the
first failed unit.  The potential for data or system availability
loss may be reduced if a spare disk unit is attached to the computer
system.  A method is described for providing the spare disk unit
capability without actually requiring that a specifically designated
spare disk unit be attached to the computer system.

      The exposure to loss of data and/or system availability due to
disk unit failure may be reduced through the use of "hot spare disk
units."  The hot spare disk units approach typically involves
attaching extra, unused, inactive disk units to the computer system
(Figs. 2a, 2b, 2c).  Then, when one of the active disk units signals
that it is about to fail (or, in the mirrored or RAID case, actually
does fail), the data stored on that unit is moved onto the spare
unit, thus avoiding the loss of data and/or loss of system
availability (or, in the mirrored or RAID case, minimizing the
duration of exposure to such loss due to the failure of a second
unit).  However, if multiple different disk unit types or sizes are
attached to the computer system, then this approach to hot spare disk
units will require attaching multiple spare disk units -- one of each
type or size.  Also, if one desires to protect against the failure of
multiple active disk units, then multiple spare disk units must be
attached.  Finally, since the attached spare disk units are inactive,
they do not contribute the usual improvement to system performance
expected when additional disk units are attached.

      The method of reducing exposure to loss of disk unit data
and/or system availability referred to as "virtual hot spare disk
units" only requires that there be adequate unused disk storage
capacity on the active disk units attached to the computer system to
contain the data from any single disk unit attached (Figs. 3a, 3b,
3c).  If the disk units on the system are grouped into logical pools,
then there must be adequate unused storage capacity to contain the
data from a single disk unit within each pool.  If, when the computer
system is powered on, there is not enough unused disk storage
capacity to hold the data from any single disk unit attached to the
system (or within each logical pool of disk units), then the system
operator will receive a warning message, but the system will continue
to run normally.

Based on these assumptions, the virtual hot spare disk units approach
works as follows:
  o  For unprotected disk units (Fig. 3a), when a dis...