Windows HPC Server 2008 Diagnostics Video
< Go Back
Diagnostics is an area in which cluster administrators spend a lot of their time. This video introduces cluster diagnostics in Windows HPC Server 2008.
The goal of cluster diagnostics is to validate a cluster after deployment or after changes in configuration, to troubleshoot any failures, and to measure any performance degradation that may occur over time. Windows HPC Server 2008 includes 16 built-in diagnostic tests. The first group of tests verifies the infrastructure, which includes scheduler, service, connectivity, and service-oriented architecture (SOA) capabilities. A second group of tests generates configuration reports focusing on applications, network topology, software updates, and services. Finally, Windows HPC Server 2008 includes several tests that measure performance bandwidth and latency in a cluster (MPIPingPong).
The video then shows how to run diagnostic tests from the Admin Console. A demonstration steps through the process of running a test; examining the results and exporting them to HTML and to a Microsoft Office Excel spreadsheet, where the results are converted to a pivot table) for further analysis; troubleshooting the problem; re-running the test to verify that the problem has been solved; and clearing the error alert.
