Browse Prior Art Database

Mechanism for Preserving Enclosure Unique Identifier when Replacing Motherboard

IP.com Disclosure Number: IPCOM000033510D
Original Publication Date: 2004-Dec-13
Included in the Prior Art Database: 2004-Dec-13
Document File: 2 page(s) / 25K

Publishing Venue

IBM

Abstract

System management software which manages multiples systems generally persists some information about each system it manages. In order to differentiate between managed systems, it may rely on some unique identifier programmed in the hardware. This unique identifier is usually stored in a persistent data device on the system motherboard. This information is programmed as part of the manufacturing process. When the planar is replaced in the field, this information needs to be reprogrammed. Currently this is accomplished as a manual step by the CE (Customer Engineer) replacing the planar. Frequently the CE forgets this step in the process. This article describes two solutions to automating the re-assignment of this unique identifier.

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

Page 1 of 2

Mechanism for Preserving Enclosure Unique Identifier when Replacing Motherboard

    This article describes a solution to this problem in an environment with a chassis or rack management device by caching the unique identifier of each system. The device must be removed from the rack or chassis to be serviced. The management device is able to determine that the system has been removed and re-inserted. When it is re-inserted into the chassis or rack, the management device will query for the persistent identifier. If this value is determined to be invalid, the management device will push the previous identifier down to the device. This provides an automated mechanism for guaranteeing that the unique identifier is programmed correctly into the replacement part.

One possible implementation of our invention is described below.

This implementation consists of the following pieces:
1) A fixed enclosure such as a bladecenter chassis
2) A management module which manages (1)
3) one or more blades which can be hosted by (1)
4) A service processor located on an instance (3) that manages it.
5) BIOS running on (3) which populates SMBIOS tables
6) System management software agent running on the managed blade
7) a centralized management server component of the system management software.
8) One or more unique identifiers for (3) programmed into a static device such as an EEPROM located on a replaceable part such as a motherboard.

     It is assumed that at the initial state the system unique identifier is correctly programmed into the motherboard on the system.

When (3) is inserted into (1), (2) will query (4) for (8). (2) will then cache this value. When (3) is powered on, (5) will execute POST and build the SMBIOS tables containing
(8). After the OS has started, (6) is assumed to be running and will query (8) from the SMBIOS table. (7) queries (6) for (8). If a representation for a system with the same value of (8) does not exist, (7) will create one.

     Some time later, (3) is removed from (1) to be serviced. The device containing the va...