Eduardo Grosclaude wrote: > On Tue, Mar 9, 2010 at 1:08 AM, John R Pierce <pierce at hogranch.com> wrote: > >> for a high performance compute cluster, you'll probably want to use >> management software like Oscar, which integrates system management with >> MPI based distributed computing such that you can manage a cluster of >> 100s of servers like its a single big system >> > > I've been using Kusu with much success. Sadly, you're pretty much on > your own there as the project seems unsupported or sucked dry out by > Platform.com. > I hope it to fully reincarnate in Red Hat's HPC proposal and that it > eventually makes its way into CentOS. > note that Oscar 6.x can be used with Centos 5.x (or debian or suse), and it seems like Centos is their preferred platform. I setup an Oscar test cluster some time ago using some old PCs, it was surprisingly easy. you install the oscar packages on your 'master' server, this one has two connections, one to your LAN and one to your HPC cluster (which is on its own switch). Then you PXE boot your HPC nodes and they get installed with a centos+scientific kit, inculding any custom application stuff you specified. then you just run your MPI based application(s), and its automatically distributed across the nodes of the cluster, Oscar also provides monitoring (Ganglia) and other stuffs. MPI is a standard Message Passing Interface used in scientific computing, essentially you write your software such that it accepts messages telling it what to do and sends messages with the results. This works best for applications that don't need a lot of global interactions, where each unit of computation can be self contained for some reasonable period of time.