On 6/17/05, Randall Shimizu <[EMAIL PROTECTED]> wrote: > > ----- Original Message ----- > From: "Randall Shimizu" <[EMAIL PROTECTED]> > To: [EMAIL PROTECTED] > Subject: mtbf and VM's ....................??? > Date: Fri, 17 Jun 2005 10:56:12 -0800
It'd be nice if you'd cut and paste just the text, and not the quotes and headers and stuff. It's an eyesore. > I was wondering if anyone has heard what the mtbf of a VM > is......??? Now I am sure there is number of variables that come > into play. A hardware based hypervisor such as ESX or IBM's rHype > will increase reliability............ What exactly does a hypervisor do? > VM failover is one feature that is still lacking in the x86 VM arena.... I would think that if you had systems that were sufficiently similar to each other that they could pick up each other's work, then they would suffer the same problems. With failover between different, non-virtual machines, you can guard against hardware-related failures. But with VMs, not only are the OS's likely to be the same or very similar, but the hardware they're running on is the same! What would cause one VM to fail, but not afflict its sibling? (It's not a rhetorical question.) And, for another question: why couldn't you apply to virtual machines the same clustering methods you'd use for non-virtuals? Wouldn't, say, openMosix work? -todd -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
