compute-0-20 is back online. Diagnosis was bad memory. Replaced under warranty, and rejoined the cluster.

compute-0-20 failed to POST during testing. It is currently being repaired under warranty, and will join the cluster as soon as it is returned.

Compute nodes now have 8G of swap, to go with the 4G of core RAM.

It took approximately 12 minutes to reload all 31 compute nodes.

Click on the links above to monitor your cluster and to learn more about how to use it.

In addition, you can optionally register this cluster. The National Science Foundation is our primary funding vehicle. While we have several measurements of Rocks usage (downloads, number of list subscribers, number of posts), the registry gives our funding agency an additional and significant device for evaluation. It also lets us know which platforms are the most important to Rocks users and what sizes of clusters our software must support.

To add more content to this site, click on the link below that is labeled: “Add content to this web site”.