NodeMon

Nodemon is a visualization tool for monitoring system resource utilization. It is used for monitoring the Columbia supercomputer, a 10,240-processor system at NASA Ames Research Center.

Nodemon uses the Growler distributed object framework.

Development

 Nodemon is distributed under the NASA Open Source Agreement (NOSA).
Copyright © 2006 United States Government as represented by the Administrator of the National Aeronautics and Space Administration. All Rights Reserved.

Authors:
Bryan Green bgreen@nas.nasa.gov
Chris Henze chenze@nas.nasa.gov (LinkMon module; cpuset hilighting in CPUMon)
Contributions are welcome. See the Source Forge projects page. SourceForge Logo

NodeMon was developed under the auspices of the Visualization Group at the NASA Advanced Supercomputing facility (NAS) at NASA Ames Research Center.
The Visualization Group members are:

Download

 Nodemon is in alpha release. The latest release is version 0.3.5 (releasenotes). Releases may be downloaded from SourceForge.net.

Nodemon needs the Growler distributed object framework. Releases of Growler may also be downloaded from SourceForge.net.

Documentation

 For the time being, there is no documentation on using Nodemon, other than in the source code. You may email me if you have questions.

Information and Links

 Nodemon was on display on a 3x3 hyperwall in the NASA booth at the SC05 conference exhibition, and was subsequently featured in the Gentoo Weekly Newsletter.
A set of images from SC05 can be found here.

Nodemon was also used at SC06, to monitor the concurrent visualization pipeline between the Columbia supercomputer and the 3x3 hyperwall in the NASA booth. Click on the image to the right for a higher-resolution image.
The top row of screens in the image show three renderers creating frames based on data being streamed in from a live simulation on Columbia. These frames are being concatenated into movies in the background. The middle row shows the corresponding movies being played in a loop. Over time, as frames accumulate, the movie grows longer and longer, detailing the entire simulation run. The third row shows a set of NodeMon clients, "NodeViews", for monitoring the pipeline. The left screen shows the state of the Columbia node running the simulation. The middle image shows the hyperwall master node, with cpu, memory, ethernet, and infiniband statistics. The right screen shows statistics for each of the 3x3 hyperwall nodes.

Here are some more screenshots, and some movies.

bgreen@nas.nasa.gov
www.nas.nasa.gov/~bgreen