[openstack-dev] [nova][scheduler] Instance Group Model and APIs - Updated document with an example request payload
Yathiraj Udupi (yudupi)
yudupi at cisco.com
Tue Oct 29 21:51:48 UTC 2013
On 10/29/13, 2:26 PM, "Chris Friesen" <chris.friesen at windriver.com> wrote:
>On 10/29/2013 03:14 PM, Andrew Laski wrote:
>
>>
> > Nova has placement concerns that extend
>> to finding a capable hypervisor for the VM that someone would like to
>> boot, and then just slightly beyond. If there are higher level
>> decisions to be made about placement decisions I think that belongs
>> outside of Nova, and then just tell Nova where to put it.
>
>Not sure about this part. I think that with instance groups Nova should
>be aware of the placement of the group as a whole, not just single
>instances.
>
>As soon as we start trying to do placement logic outside of Nova it
>becomes trickier to deal with race conditions when competing against
>other API users trying to acquire resources at the same time.
APIs are designed to be 2-stage as per this proposal. First register the
entire instance group, and then trigger another api to create them. These
create calls can be scheduled one by one if necessary.
Agreed that these creates will create race conditions, hence it definitely
needs some concept of a reservation to be really accurate. It is not
enough to just find the ideal optimal resource placement at any point of
time, there should also be guarantee that it can be eventually scheduled.
Hence the actual implementation of these apis, will need such additional
support.
Thanks,
Yathi.
More information about the OpenStack-dev
mailing list