Browse Prior Art Database

3990 RAMAC Recovery Copy

IP.com Disclosure Number: IPCOM000115786D
Original Publication Date: 1995-Jun-01
Included in the Prior Art Database: 2005-Mar-30
Document File: 2 page(s) / 78K

Publishing Venue

IBM

Related People

Beardsley, BC: AUTHOR [+8]

Abstract

The RAMAC Recovery Copy function automatically copies all data from a failed or failing RAMAC drawer to an available spare drawer, provided either one HDA or the Cache Manager, but not both, encounter a failure. RAID 5 algorithms and Dual Copy algorithms are utilized to provide uninterrupted access to all of the data in the drawer. Once all of the data has been copied to the spare drawer, the failing drawer can be repaired or replaced so that it is available to have its original data restored or it can function as a spare in case of another drawer failure.

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

3990 RAMAC Recovery Copy

      The RAMAC Recovery Copy function automatically copies all data
from a failed or failing RAMAC drawer to an available spare drawer,
provided either one HDA or the Cache Manager, but not both, encounter
a failure.  RAID 5 algorithms and Dual Copy algorithms are utilized
to provide uninterrupted access to all of the data in the drawer.
Once all of the data has been copied to the spare drawer, the failing
drawer can be repaired or replaced so that it is available to have
its original data restored or it can function as a spare in case of
another drawer failure.

      A RAMAC device consists of a rack of drawers, each drawer
containing Head-Disk Assembly (HDAs) and a Cache Manager (an
electronic memory cache and a microprocessor to manage the cache
allocation, staging and destaging, etc.).  Some of the drawers may be
allocated as spare drawers for recovery purposes.

      RAMAC Recovery Copy deals with the recovery of the 3390 logical
volumes when a RAMAC drawer fails, either due to a HDA failure or to
a RAMAC Cache Manager failure.  A drawer can be recovered if one HDA
is broken or if the Cache Manager is broken.  If more than one HDA is
broken, the affected volumes in the drawer cannot be recovered.  The
recovery is achieved by copying the volumes from the broken drawer to
the volumes of the spare drawer.  The RAID 5 algorithm is central to
the recovery process.  Damaged data is regenerated through use of
parity and valid data track bands.  Out of sync parity is likewise
regenerated from the data bands (i.e., from the relevant data
tracks).

      Spare drawers are distinguished from customer use drawers by
additional bit settings on the device status track.  These bits are
set at device install time and can be modified at any time
thereafter, except when recovery copy is in progress to the drawer.
The spare indicators are set on the physical device rather than the
control unit to allow for sp...