International Business Machines Corporation

Hany A Salem

Leigh Allen Williamson

Donald Alexander Bourne

Michael J Casile

Ajay Arvind Apte

James William Stopyro

Charles James Redlin

Francis Gerard Malin

Problem diagnostics may be obtained from components that log messages using a unique component identifier which is provided within each message so as to enable routing of callbacks to the component that originated the message. Similarly, problem diagnostics may be obtained from components that generate alerts, where the alerts contain the unique identifier of the component that generated the alert. Each component supports a diagnostic provider interface comprising operations which may be dynamically invoked, for example to solicit information from the component such as its configuration data, its state information, to execute self-diagnostic tests, and so forth. In addition or instead, operations may be provided that can be invoked to cause the component to alter its configuration data (such as notifying the component to change its level of message logging) or its behavior.

Closed-loop diagnostic system

A solution provided here comprises transferring control to at least one data-capture component when an error occurs in a production environment, performing introspection upon an object associated with the error, and capturing data from the object. The captured data is associated with the error, and the capturing is performed at least partially through introspection. Methods for error&#x2014;handling and providing diagnostic data, systems for executing such methods, and instructions on a computer-usable medium, for executing such methods, are provided.

Object introspection for first failure data capture

Closed-loop diagnostic method

A method, apparatus, system, and signal-bearing medium that in an embodiment determine whether to log an exception based on whether the exception has already occurred. In an embodiment, exceptions that repeatedly occur at the same source and location within the source are logged on the first occurrence, and not logged thereafter. In this way, data can be saved about the exception, yet system resources are not consumed by repeated logging.

Logging of exception data

An example of a solution provided here comprises: selecting an incident that requires further processing, capturing data associated with said incident, outputting captured data, and outputting a description for said captured data, whereby problem-solving is promoted. Another example comprises: providing runtime features for data capture, selecting an incident that requires further processing, capturing data on a thread that encounters said incident, and outputting captured data. In some cases, such a solution might include comparing a current incident to known incidents, and if a match is found, retrieving information that is relevant to said current incident. In some cases, such a solution might include taking recovery action or corrective action in response to said incident. Methods for handling errors, systems for executing such methods, and instructions on a computer-usable medium, for executing such methods, are provided.

First failure data capture

An example of a solution provided here comprises analyzing an incident, responding to said incident, and customizing the response, based on the result of the analysis. Another example of a solution comprises accessing a knowledge base, comparing a current incident to known incidents in said knowledge base, and if a match is found, retrieving information from said knowledge base, said information being relevant to the current incident. Methods for handling errors, systems for executing such methods, and instructions on a computer-usable medium, for executing such methods, are provided.

Growing community of inventors

Pflugerville, TX, United States of America

Hany A Salem