Rocks

Contacts

Rocks is used at CS and CAC.

• 3/13/13: Lulu and Oliver met with CS's Martin Berggren <mjb43>, to discuss Rocks.
• Oliver can make introductions to CAC staff, including Resa Alford <resa.alvord> and Steven Lee, among others.
  • CAC runs Nandini's cluster. Specs to its OS and provisioning software are available here:
    • http://www.cac.cornell.edu/wiki/index.php?title=ASTRA_Cluster

Resources

11/28/12 email to Lulu, from Oliver:

Here’s what appears to be a good overview in setting up an HPC cluster. This is the 2nd of 6 “pages”, which has a link to Rocks <http://www.rocksclusters.org>, but not Warewulf:

HPC ToolboxHPC Toolbox; Dominic Eschweiler

• http://www.admin-magazine.com/HPC/Articles/Exploring-the-HPC-Toolbox/

4/25/14 email to Lulu, from Oliver:

I parked Martin's info below (in his use of SSSD, though he doesn't mention it specifically) at our ever-growing wiki page about Linux and CU AD:

https://confluence.cornell.edu/x/PQdIDg

As you know, he uses Puppet. And his clusters are Rocks; as I shared with you, each of his compute nodes authenticates (and authorizes?) against CU AD.  And although the CIT project is in a "pilot", he is using their service in full-bore production mode for all his systems. :-)

Warewulf is what ChemIT uses

These articles on about using Warewulf  <http://warewulf.lbl.gov/trac>, as emailed to Lulu, from Oliver, 11/28/12:

Warewulf Cluster Manager – Howlingly Great; By Jeff Layton

• http://hpc.admin-magazine.com/Articles/Warewulf-Cluster-Manager-Master-and-Compute-Nodes

Warewulf Cluster Manager – Part 2; By Jeff Layton

• http://hpc.admin-magazine.com/Articles/Warewulf-Cluster-Manager-Completing-the-Environment

Warewulf Cluster Manager – Part 3; By Jeff Layton

• http://hpc.admin-magazine.com/Articles/Warewulf-Cluster-Manager-Development-and-Run-Time

Warewulf Cluster Manager – Part 4; By Jeff Layton

• http://hpc.admin-magazine.com/Articles/Warewulf-Cluster-Manager-Administration-and-Monitoring

Other ideas

Michael Hint noted 6/26/14 that is seems likely that someone in the Engineering college uses OpenStack (for clusters?):

• http://www.openstack.org/

Oliver found this article on "wetting up a real-world HPC cluster with Kickstart, SSH, Son of Grid Engine, and other free tools", by a SysAdmin, Gavin W. Burris, at the University of Pennsylvania School of Arts and Sciences:

• http://www.admin-magazine.com/HPC/Articles/real_world_hpc_setting_up_an_hpc_cluster
• His blog, <http://idolinux.blogspot.com/> has an HPC in AWS example, but is otherwise dated:
  • http://idolinux.blogspot.com/2012/06/fabcluster-amazon-ec2-script.html

Another resource Oliver found, 11/13/14:

A cluster monitoring tool suite. "Wulfware: A LAN/Cluster/Beowulf Monitoring Suite; Version 2.6.0; Robert G. Brown (rgb).

• This is the official website for the Wulfware suite of LAN or cluster monitoring tools:
  • http://www.phy.duke.edu/~rgb/Beowulf/wulfware.php