URL:
Reporter RHNID:
Section: -
Language: en-US (English)
Workaround:

Description: Chapter 7: Troubleshooting Compute Nodes Narrative
Rodolfo: I’d like to ask why we have a very short explanation about regions, az’s, aggr hosts and cells. It’s quite important. Please read https://www.openstack.org/assets/presentation-media/divideandconquer-2.pdf. Amazon gives this feature and I think it’s very interesting. I think is a day 1 issue for operator. Where to place instances depending on the kind of workload, affinity for low latency, etc, etc.These kind of “tuning” make things much faster. Specially in global clouds where you want to choose regions near to you for latency reasons or different az for instances in same project for redundancy and HA. Please read https://cloud.google.com/compute/docs/regions-zones/regions-zones
This concepts are slightly introduced in chapter 6 talking about migrations. Maybe there we could insert some diagrams supporting these ideas.