On Thu, Oct 13, 2011 at 3:57 PM, Lachele Foley (Lists) <[email protected]> wrote: > I have two servers. They're not set up yet, but I plan to use them in > failover. Half of the clients will be assigned to one machine, and > half to the other. I'll use automated deployment tools (chef & > vagrant, most likely) to maintain the machines. I'll use those same > tools to maintain essentially identical(*) virtual images (virtualbox) > of each machine, but on the other host. So, if one machine goes down, > the other machine will boot up a VM of the dead machine and that VM > will do the terminal service while the real machine is repaired. This > means that if one machine dies, everyone has to operate on half the > normal resources, but at least things don't stop altogether. > > (*) The VM will look just like the normal machine in most ways, but > I'll include some messages here and there so users will know if > they're booting from the VM rather than the normal setup. > > I'm sure there are other ways to do what you want, too.
That's a really interesting idea. In the meantime I realised that the tftp server is read only, which makes failover so much easier =) I,.e. you don't have to start/stop the service, you can just let tftp run on both machines and use ucarp (or heartbeat). http://www.pytips.com/2010/3/16/ip-failover-with-ucarp-on-ubuntu ucarp is easier to use than heartbeat. ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2d-oct _____________________________________________________________________ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.freenode.net
