Would an "initializing" state for a VM that exists in the hypervisor but is not completely set up (e.g., cannot be started, yet) be useful? This may be somewhat like "blocked", but it's a specific case that may be worth distinguishing because the available information about the VM may be incomplete.
Thanks, --David > -----Original Message----- > From: [email protected] [mailto:[email protected]] On Behalf Of > Michael MacFaden > Sent: Thursday, March 07, 2013 1:09 PM > To: Juergen Schoenwaelder > Cc: opsawg > Subject: Re: [OPSAWG] VMM-MIB joint proposal: Item 3 - VM-MIB paused != > suspended > > > OK. If we add an administrative suspend state, we likely also need a > > corresponding operational state, no? > > Agreed. A VM would transition to paused(5) state imo. > > VirtualMachineOperState ::= TEXTUAL-CONVENTION > STATUS current > DESCRIPTION > "The operational state of a virtual machine: > > unknown(1) The state is unknown, e.g., because the > implementation failed to obtain the state > from the hypervisor. > > other(2) The state has been obtained but it is > not a known state. > > running(3) The virtual machine is currently running. > > blocked(4) The virtual machine is currently blocked. > > paused(5) The virtual machine is currently paused. > > migrating(6) The virtual machine is currently > migrating. > > shutdown(7) The virtual machine is currently in the > process of shutting down. > > shutoff(8) The virtual machine is down. > > crashed(9) The virtual machine has crashed." > > > Now if admin state is set to suspend, then the resting state of the VM > I believe should be shutoff(7) if operation was successful otherwise > the VM transitions to one of running, blocked, crashed, other, unknown. > > Given suspend operations can take a long time some implementations > may report the transition to shutdown(7) before shutoff(7) while other > implementations will not. > Existing fielded IF-MIB implementations exhibit this behavior as well. > While IF-MIB module provides a rich set of states, many implementations > only use a subset. > > Will send out more text to make explicit what we the designers > intend was for state transitions. > > Mike MacFaden > _______________________________________________ > OPSAWG mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/opsawg _______________________________________________ OPSAWG mailing list [email protected] https://www.ietf.org/mailman/listinfo/opsawg
