Support for multiple active scheduler drivers

Registered by Alex Glikson

In heterogenous environments, it might be desirable to apply different scheduling policies in different host aggregates. This could be different drivers, or even same driver with different configurations (e.g., FilterScheduler with different sets of filters/weights and/or different parameters of particular filters/weights).

[Details in the wiki have been updated to reflect the proposed design for Icehouse]

Blueprint information

Status:
Started
Approver:
Russell Bryant
Priority:
Low
Drafter:
Alex Glikson
Direction:
Needs approval
Assignee:
Inbar Shapira
Definition:
Drafting
Series goal:
None
Implementation:
Needs Code Review
Milestone target:
None
Started by
Alex Glikson

Related branches

Sprints

Whiteboard

Gerrit topic: https://review.openstack.org/#q,topic:bp/multiple-scheduler-drivers,n,z

Addressed by: https://review.openstack.org/58480
    Support multiple active scheduler policies/configurations

Addressed by: https://review.openstack.org/61794
    conf overrides in scheduler

Addressed by: https://review.openstack.org/61821
    Leverage unified conf overrides in scheduler

Addressed by: https://review.openstack.org/62767
    Add db api to get aggregate by key value

Its a bit late in Icehouse to merge this, the performance risk seems really high. We need to cut some blueprints, and I am afraid this one has been chosen. Certainly worth pushing back up when Juno opens. Need to take some care about not affecting the peformance of the faster, but less feature-full path --johnthetubaguy

Unapproved - please re-submit via nova-spec --johnthetubagy (20th March 2014)Gerrit topic: https://review.openstack.org/#q,topic:bp/multiple-scheduler-drivers,n,z

Addressed by: https://review.openstack.org/58480
    Support multiple active scheduler policies/configurations

Addressed by: https://review.openstack.org/61794
    conf overrides in scheduler

Addressed by: https://review.openstack.org/61821
    Leverage unified conf overrides in scheduler

Addressed by: https://review.openstack.org/62767
    Add db api to get aggregate by key value

Its a bit late in Icehouse to merge this, the performance risk seems really high. We need to cut some blueprints, and I am afraid this one has been chosen. Certainly worth pushing back up when Juno opens. Need to take some care about not affecting the peformance of the faster, but less feature-full path --johnthetubaguy

Unapproved - please re-submit via nova-spec --johnthetubagy (20th March 2014)

Marking this blueprint as definition: Drafting. If you are still working on this, please re-submit via nova-specs. If not, please mark as obsolete, and add a quick comment to describe why. --johnthetubaguy (20th April 2014)

(?)

Work Items

Work items:
support for overrides of scheduler config options based on availability zone: INPROGRESS
Update docs: TODO
support overrides of scheduler_driver and scheduler_host_manager: POSTPONED
support policy selection based on tenant: POSTPONED