[plug] VM made out of a cluster of real servers?
meyerri at gardenshark.org
Sun Jul 1 09:49:25 WST 2012
I can give you a bit of info on the mainframe, after all I've been
working off and on with them since the 70's (Yes, I AM that old).
No operating system (as far as I know, and I'm pretty knowledgeable,
having worked with MVS+AF8-OS/390+AF8-z/OS, VM/SP+AF8-VM/HPO+AF8-VM/XA+AF8-VM/ESA+AF8-z/VM and
VSE in it's many manifestations - no z/Linux (yet)) will actually run an
image across multiple machines. You can Sysplex multiple instances of of
z/OS across multiple machines, be they physical separate machines,
physical partitions of the same machine or separate virtual machines
under VM. They are not the same image and cannot just run a workload
spread out like that. The sysplex enables workloads to be moved from one
machine to another in case of work load manager requirements or a
machine being shut down, but the whole is still less than the sum of its
parts as they say.
And, yes, processors can be taken offline and replaced with no downtime.
As for removing and replacing any particular processor from any VM, I'm
not sure - the VM may be bound to a single particular processor ... what
happens if that has to be offlined? Hmmm .... I don't know.
On Fri, 2012-06-29 at 15:49 +-0800, Andrew Cooks wrote:
+AD4 Consider the IBM mainframe. It was around before the beginning of time
+AD4 (00:00:00 UTC on 1 January 1970). It can run multiple OSs. It has
+AD4 multiple independent processor boards that you can replace without
+AD4 shutting down the whole machine. Can any arbitrary processor be
+AD4 assigned to, or removed from any VM?
+ACI-He who fights with monsters should look to it that he himself
does not become a monster...when you gaze long into the abyss,
the abyss also gazes into you.+ACI
(Friedrich Nietzsche, 1878)
Linux Counter user +ACM-306629
More information about the plug