Browse Prior Art Database

Ensuring Data Availability on a Redundant Array of Independent Disks Box while Plugging in a New Drive

IP.com Disclosure Number: IPCOM000113892D
Original Publication Date: 1994-Oct-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 2 page(s) / 42K

Publishing Venue

IBM

Related People

Crews, C: AUTHOR [+6]

Abstract

Described is a method for allowing a new drive to be plugged into a Redundant Array of Independent Disks (RAID) box without disrupting the data transfers with other drives that share the same Small Computer System Interface (SCSI) cable.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 76% of the total text.

Ensuring Data Availability on a Redundant Array of Independent Disks
Box while Plugging in a New Drive

      Described is a method for allowing a new drive to be plugged
into a Redundant Array of Independent Disks (RAID) box without
disrupting the data transfers with other drives that share the same
Small Computer System Interface (SCSI) cable.

      When the controller card receives a command that involves
turning on the power to a new drive and spinning it up, the microcode
needs to finish up all of the ops that are currently ongoing with any
of the other drives that share the same SCSI bus.  Once all of the
activity on the SCSI bus has stopped, the controller card leaves all
the drives on that bus in this quiesced mode by holding off any
requests that would involve SCSI activity.  Note:  it is important to
stay in this mode as briefly as possible in order to avoid having the
host timeout.  After the bus has been quiesced, the power envelope to
the new drive can be turned on.  As soon as it is up the SCSI spin-up
command is issued and a short timer is started.  The duration of the
timer should be slightly longer than it will take the SCSI bus to
return with a communication failure if the new drive does not have
the correct ID, but short enough that the host will not timeout.  If
the new drive has the correct ID and is in the process of spinning
up, then the timer will expire prior to getting response from the
drive.  At this point the other drives on the...