How to Troubleshoot a Server Abend

NetWare servers tend to abend when they get unhappy. This can happen when

• a server runs out of resources or
• a poorly written application does not properly manage resources, which causes an operating system fault that causes the application to be terminated.

This happened quite often with GroupWise Web Access document viewer in versions of GroupWise prior to 7.0.

Getting Ready to Troubleshoot

To troubleshoot this issue, more information about the actual cause of the error must be gathered. Before attempting to reproduce the issue, debug / trace logging and network monitoring needs to be configured, so that the exact point of failure can be captured.

Riva Application Logs

Riva has several logging levels. These levels limit the amount of information logged to reduce disk (IO) usage and to conserve disk space.

To troubleshoot:

1. Enable DEBUG Logging level.
2. Reduce the execution thread counts to 1 or a single policy at a time.

GroupWise POA SOAP Tracing

The GroupWise server has a capability called GWTRACE that will log the SOAP XML request and response directly to the Post Office log directory.

Riva can enable this tracing capability.

To enable the tracing capability:

• In the root of the Riva folder, create a blank file with no extension named DEBUG.

Micro Focus, the new owners of GroupWise, have reorganized the GroupWise documentation. The two links we used to have here have been redirected to a single page that merely lists various SDK guides. For documentation on GWTRACE or GroupWise XML, try GroupWise SDK Guides,

HTTP Communication Monitoring

To capture the actual network data being sent and received between Riva and the GroupWise Post Office, the network communication must be intercepted. The best tool to do this is an HTTP Proxy service that can be used to inspect and record HTTP communication.

At this point, lower-level tools that allow for packet level inspection like Wireshark are not yet needed.

The preferred tool by Riva developers and support is Fiddler.

Download here: https://www.telerik.com/download/fiddler

Finding the Underlying Cause

The key to being able to resolving the abend is finding the root cause by being able to systematically reproduce the problem.

With all the monitoring and different components logging in diagnostics. 

1. Clear all logs and start all logging from scratch.

2. Start the process that is causing the abend.

3. Reproduce the abend.

4. Gather the logs and network traces into a separate folder and provide applicable screen shots.

5. Review for any insights into the cause, and try to reduce the number of logs by pin-pointing a specific time frame.

6. Start at Step 1 and repeat until the specific event / API call causing the error has been identified.