Ask Your Question

Does OpenStack dynamically scale or migrate instances when compute node becomes overloaded?

asked 2014-10-24 03:01:28 -0500

Daneel gravatar image

I'm beginner in OpenStack and trying to study and understand its principles. During considerations if and how to implement PoC of OpenStack we've wondered if there is a something like RHEV Evenly Distributed or vmware DSR. I've tried to search through OpenStack Architecture Design Guide and other documents but haven't got direct answer.

In other words, when several instances on a compute node "get crazy" and overload the node will controller somehow move some instances away from this compute node to balance workload automatically?


edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2014-10-24 17:51:12 -0500

jtaguinerd gravatar image

HI, Afaik there's no automagical way of migrating the instance, unless of course you have a third party tool that monitors the load of a particular compute node and provided you have a shared storage that will enable you to transfer disk of the instance. What the controller will do, particularly the scheduler, is to stop launching instance in an overloaded compute node and passed on the request to the another nodes that has available resources.


edit flag offensive delete link more


Thanks for your answer. Are there any references in documentation on behaviour you are describing?

Daneel gravatar imageDaneel ( 2014-10-30 11:13:04 -0500 )edit

You can further read up scheduling from here

jtaguinerd gravatar imagejtaguinerd ( 2014-10-30 21:32:42 -0500 )edit

Get to know Ask OpenStack

Resources for moderators

Question Tools



Asked: 2014-10-24 03:00:21 -0500

Seen: 858 times

Last updated: Oct 24 '14