Correlated Diagnostic Data in Incident Packages

To improve the diagnosability of a problem, it is sometimes necessary to examine not only diagnostic data that is directly related to the problem, but also diagnostic data that is correlated with the directly related data. Diagnostic data can be correlated by time, by process ID, or by other criteria. For example, when examining an incident, it may be helpful to also examine an incident that occurred five minutes after the original incident. Similarly, while it is clear that the diagnostic data for an incident should include the trace file for the Oracle Database process that was running when the incident occurred, it might be helpful to also include trace files for other processes that are related to the original process.

Thus, when problems and their associated incidents are added to a package, any correlated incidents are added at the same time, with their associated trace files.

During the process of creating the physical file for a package, the Support Workbench calls upon the Incident Packaging Service to finalize the package. Finalizing means adding to the package any additional trace files that are correlated by time to incidents in the package, and adding other diagnostic information such as the alert log, health checker reports, SQL test cases, configuration information, and so on. This means that the number of files in the zip file may be greater than the number of files that the Support Workbench had previously displayed as the package contents.

The Incident Packaging Service follows a set of rules to determine the trace files in the ADR that are correlated to existing package data. You can modify some of those rules in the Incident Packaging Configuration page in Enterprise Manager.

Because both initial package data and added correlated data may contain sensitive information, it is important to have an opportunity to remove or edit files that contain this information before uploading to Oracle Support. For this reason, the Support Workbench enables you to run a command that finalizes the package as a separate operation. After manually finalizing a package, you can examine the package contents, remove or edit files, and then generate and upload a zip file.

Note:
Finalizing a package does not mean closing it to further modifications. You can continue to add diagnostic data to a finalized package. You can also finalize the same package multiple times. Each time that you finalize, any new correlated data is added.