<div class="gmail_quote">On Wed, Aug 24, 2011 at 4:20 PM, Jomegat <<a href="mailto:jomegat@jomegat.com">jomegat@jomegat.com</a>> wrote: <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"> <div class="im"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> For example, I don't have the hardware here to be able to test LTSP cluster capability, so I may pay LTSP upstream to implement, test and document LTSP cluster for K12Linux. </blockquote> </div> What do you mean by "LTSP cluster capability?" Multiple terminal servers? What would you need to gain that capability? </blockquote><div> Think about a number of servers in a closet that are ready to answer a PXE boot request and join any client to the network that asks. There's a couple of ways I was approaching this post APS-project: 1. multiple servers providing the DHCP/PXE capabilities with no load balancing at all. So which ever server answered first would take the client (only works in pairs and is rather a network noise problem). 2. multiple LTSP servers with a single (or failover pair of) DHCP/PXE server(s) that would load balance based on feedback from the LTSP servers. This would allow for a much larger stack of servers for the LTSP side and the ability to do enterprise things like NAS/SAN/Fiber Channel storage, etc. Also allows for hardware replacement with minimal downtime for clients - tell dhcp server to not use a system then send a reboot to the clients (in between classes). 3. Cluster the servers using a "big" tool like OpenSSI, Mosix, or Kerrighed. This is the opposite of taking a big machine and splitting it out into smaller virtual machines. It takes a bunch of physical machines and creates a single, large virtual machine with many processors and lots of ram. The cat's meow of this is the ability to migrate processes from a failing hardware node in order to fix it. Additionally, physical nodes that are not maxed out can take load automatically to do node-wise load balancing. Plus it's way cool for kids to be able to say they use a supercomputer every day at school :-) </div></div>-- -- James P. Kinney III As long as the general population is passive, apathetic, diverted to consumerism or hatred of the vulnerable, then the powerful can do as they please, and those who survive will be left to contemplate the outcome. - 2011 Noam Chomsky <a href="http://heretothereideas.blogspot.com/" target="_blank">http://heretothereideas.blogspot.com/</a>