Hi everyone,

I have currently two servers A and B with the Newton version of Openstack.

On A, I have the "basic installation" of Openstack and now i want to add a new compute node B. I want to know if it's possible to configure nova-scheduler to fairly allocate the instance on the hosts.

For example, on A, there are 2 instances using 4 VCPU and 4 Go ram and on B, there is only 1 instance using 2 VCPU and 2Go ram, nova-scheduler should create the new instance on the host of B. However, in my current setup, nova-scheduler fill the host A before using the host B.

I have followed the guide of Openstack step by step so I don't know why it works like this. I tried to use filters to configure nova-scheduler to allocate the instance based on the number of CPU available (CoreFilter) but it doesn't change anything.

Any idea? Thanks in advance.