Browse Prior Art Database

A Database and Subscription Mechanism to Minimize Time Running Hardware Diagnostics Disclosure Number: IPCOM000199812D
Publication Date: 2010-Sep-16
Document File: 2 page(s) / 29K

Publishing Venue

The Prior Art Database


This article describes a process for using a database and subscription mechanism for minimizing the amount of time a computer system spends running hardware diagnostic procedures during Initial Program Load (IPL).

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 2

A Database and Subscription Mechanism to Minimize Time Running Hardware Diagnostics

Disclosed is an autonomic mechanism wherein the system determines what diagnostics to run on which hardware. The advantages of this solution are:

Simplicity: From a customer standpoint, they do not have to worry about IPL modes and training operators on when/how to use them
Performance: Only appropriate diagnostics are performed and only on a small subset of the hardware that is determined to need it

Performing hardware diagnostics at start up is a vital way to ensure systems will perform reliably, but on large server systems with a huge number of intricately interconnected parts it is also very time consuming. In order to keep this time reasonable it is important to only run full diagnostics when appropriate. In the past this has been done primarily with "mode settings" (e.g. "fast" and "slow" IPL modes), but these are not granular (they are "all or none") and they require a human operator to make the determination which mode is appropriate at which times.

The core of the mechanism here consists of the following in the service processor that controls running hardware diagnostics on the main Central Electronic Complex (CEC) hardware:

A persistent database which maintains a list of "marked" hardware and information about which hardware/components have not yet completed required diagnostics
A subscription list which identifies which components of the diagnostics firmware require knowledge of which types of changed hardware

"Marked" hardware indicates hardware on which one may want to run some/all diagnostics. This includes:

New hardware added to the system

Hardware which has been reseated/reconnected to the system

Hardware on which a runtime problem was discovered

This results in the following flow in the service processor firmware:

Before IPL of the CEC, a hardware p...