Browse Prior Art Database

A method to build a local high capacity storage within a compact server blade chassis

IP.com Disclosure Number: IPCOM000012037D
Original Publication Date: 2003-Apr-03
Included in the Prior Art Database: 2003-Apr-03
Document File: 1 page(s) / 39K

Publishing Venue

IBM

Abstract

The blades or individual servers installed in a server blade chassis provide for high density server function. The fact that the individual blades are inexpensive and can be easily hot plugged provides for some level of redundancy and fault tolerance. If a blade malfunctions, the processing power can be easily replaced by inserting another blade. However, data stored on the malfunctioning blade's hard disks are lost. In addition, the blade architecture doesn't allow for storage expansion or fault tolerance, for example, it is not possible to easily add a RAID subsystem to a blade chassis. This invention provides for the recovery of lost data on a blade if it malfunctions.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 56% of the total text.

Page 1 of 1

  A method to build a local high capacity storage within a compact server blade chassis

       Disclosed herein is a method and system to provide data redundancy to a server blade environment. A server blade chassis typically houses about 15 server blades. Each blade, an individual server, has one to two onboard hard disk drives. Typically, one hard drive contains about 300 MB of operating system and application code and the second hard drive is unused by the system. This leaves about 35 to 70 GigaBytes of unused hard disk space per blade. By dedicating or using one blade as a file system server, all the blades in the chassis can build up a storage pool topping one-tera bytes of disk storage by contributing all unused disk storage to the common pool controlled by the file system. The mid-plane between blade server, which serves as a high speed bus, permits very high speed data transfer between blades which is much faster than a typical networked storage environment.

     By pooling all unused hard disk storage on a blade, one blade can be dedicated to run a file subsystem server application along with other applications to control the unused hard disk storage located throughout the chassis as though it is a single file system such as a RAID system. If the pooled file subsystem is treated like a Raid 5 storage array, striping and use of parity data will allow the recovery of a single drive outage. The file subsystem can also use a mirroring technique, where the second copy is s...