Trace File Analyzer Collector – A targeted diagnostic data collection utility for systems running Oracle software

Map Unavailable

Date(s) - 28/03/2013
12:00 am UTC

Live Event
Not available
Bill Burton, Sandesh Rao


Abstract: One of the biggest challenges for customers and support engineers working Grid Infrastructure (GI) and Real Applications Clusters (RAC) issues is collecting large amounts of pertinent data in a timely fashion across multiple nodes in a cluster. A big part of this challenge is frequently the data that is required is lost or overwritten, due to the diagnostic collection not happening until sometime after the problem occurred. For single Instance databases Automatic Diagnostic Repository (ADR) does a good job of generating trace packages when an incident occurs, but ADR does not include clusterware trace files and is not RAC aware. For the clusterware the diagcollection tool ( gathers together all the logs when manually requested to do so, and though this tool has proved to be useful, it is run sometime after the problem has occurred and is generally run to gather all possible logs without regard to their relevance which can make the data set very large. Further it has to be run across all nodes of a cluster which makes it more difficult to manage from an operational perspective. TFA Collector (from this point referred to as TFA) overcomes some of these problems by running a diagnostic JVM on each node in the cluster synchronized to determine when diagnostic collection is required as well as collecting and extracting only the data that is pertinent to resolving the issue. TFA is written outside of the GI /RAC/RDBMS product lines and as such could be used for any trace data, and is version and platform agnostic. In this presentation we will cover cluster wide installation, important features and common commands to do cluster wide diagnostic data collection using Trace file analyzer collector tool.


Leave a Reply