Browse Prior Art Database

System and Method for Restoring Diskless Server Blade Operations using Hibernation Files

IP.com Disclosure Number: IPCOM000012033D
Original Publication Date: 2003-Apr-03
Included in the Prior Art Database: 2003-Apr-03
Document File: 2 page(s) / 43K

Publishing Venue

IBM

Abstract

This invention utilizes remote storage area network (SAN), the hibernate sleep state and a management module to provide a very fast recovery and initialization of a blade server

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

Page 1 of 2

  System and Method for Restoring Diskless Server Blade Operations using Hibernation Files

       This disclosure pertains to server blade environment, specifically the future diskless server blades. Diskless server blades are housed in a chassis. They share the same environment, such as backplane, power supplies, network switches, and management modules. They have the same hardware components although in different slots of the chassis. Most important of all they do not have local hard disk drives. Their storage is typically at a remote location and uses shared access of operating system and applications via storage area network. As a result, other than VPD, slot location differences and IP address, such diskless blades are identical to each other. When a new blade is brought on-line (due to failure of an old blade, initial installation or an upgrade) it means plugging in a new blade, setting up the new blade, reinstalling the operating system, and restarting all of the services and applications that are needed. What is needed is a quick and easy method for bringing a new blade online and running the applications needed by the customer.

     Our invention utilizes remote storage area network (SAN), the hibernate sleep state and a management module to provide a very fast recovery and initialization of a blade server. The description below will go over a preferred implementation as well as discuss the parts of the implementation that we feel would pertain to the patentability of this disclosure.

     A typical installation of a blade center would involve a system administrator bringing up one of the blades with the desired operating system and install the desired applications. The system administrator would then force the system into a replication mode where the system automatically is put into hibernate by the management module to create a system image on the remote storage node (the hibernation file, the operating system, and installed software) that may be used by other blades in the system. The administrator selects other local blades to receive the system image (possibly equal to the number of blades in the blade center plus one or two additional backup images). The management module would update its table to associate each blade in the blade center with a system image. At this point, we now have an entire blade center full of backup blades to the original blade. A customer may decide that this is the desired redundancy of operation required, and the original blade would be brought out of hibernation to begin normal operation. At a later time, any failure to this blade's hardware and one of the replication blades could be selected to be instantly on-line as its replacement.

     A more efficient use of the blade center resources would be to add additional steps to the above configuration to provide several operational blades with different IP addresses running the same operating system and software as the original...