Placement API claiming resources
For the moment, the Placement API only calculates the free resources by decreasing the inventories with the sum of allocations. That is suboptimal because in the meantime where a request is placed and a request is fulfilled, other requests can come in and getting a wrong target in case the resource usage wasn't accurate.
The idea of that blueprint is to allow the Placement to have a way to consider intermediate resource locking until the request is fully fulfilled (ie. an instance is created).
Blueprint information
- Status:
- Complete
- Approver:
- Matt Riedemann
- Priority:
- Essential
- Drafter:
- Sylvain Bauza
- Direction:
- Approved
- Assignee:
- Jay Pipes
- Definition:
- Approved
- Series goal:
- Accepted for pike
- Implementation:
-
Implemented
- Milestone target:
-
pike-3
- Started by
- Matt Riedemann
- Completed by
- Matt Riedemann
Related branches
Related bugs
Sprints
Whiteboard
Gerrit topic: https:/
Addressed by: https:/
Super WIP: Placement doing claims
Approved for Pike. Let's get started on POC so we have something to talk about at the Forum in Boston. -- mriedem 20170425
Addressed by: https:/
Add reserved_host_cpus option
Addressed by: https:/
WIP: Don't delete allocation if instance being scheduled
Addressed by: https:/
WIP: Delete existing allocations for move operations or retries
Addressed by: https:/
WIP: Allocate instance by the conductor
Addressed by: https:/
Add a convenience attribute for reportclient
Addressed by: https:/
WIP - pass all instance IDs to scheduler
Addressed by: https:/
WIP - have scheduler claim resources
Addressed by: https:/
Make the method to put allocations public
Addressed by: https:/
Pass a list of instance UUIDs to scheduler
Addressed by: https:/
Remove selecting host based on host_subset_size
Addressed by: https:/
WIP: Scheduler returning more than 1 host
Addressed by: https:/
WIP - Pass the number of hosts to return explicitly
Addressed by: https:/
WIP - return multiple hosts from select_destinations
Addressed by: https:/
Fix ChanceScheduler returning a wrong node
Addressed by: https:/
_schedule_
Gerrit topic: https:/
Addressed by: https:/
scheduler: alternates for filter scheduler
Addressed by: https:/
placement: add claim_resources() to report client
Addressed by: https:/
placement: claim resources in filter scheduler
Addressed by: https:/
scheduler: isolate _get_sorted_hosts()
Addressed by: https:/
add dict of allocation requests to select_dests()
Addressed by: https:/
claim resources in placement API during schedule()
Addressed by: https:/
Handle None returned from get_allocation_
Addressed by: https:/
placement: add retry tight loop claim_resources()
Addressed by: https:/
WIP - add alternate hosts
Addressed by: https:/
WIP - return alternates along with their allocations
Addressed by: https:/
Set IronicNodeState
Addressed by: https:/
add a retry on DBDeadlock to _set_allocations()
Addressed by: https:/
docstring and unused code removal
Addressed by: https:/
placement: account for move operations in claim
Addressed by: https:/
Test resize with placement api
Addressed by: https:/
Sanity check delete_
Addressed by: https:/
DNM: trying to turn on periodic tasks
Addressed by: https:/
DNM: simulate rpc delay when cleaning up allocations from source node
As of https:/
Gerrit topic: https:/
Addressed by: https:/
Require Placement 1.0 in nova-status upgrade check
Gerrit topic: https:/
Addressed by: https:/
WIP - Add allocations to the values returned from the scheduler
Gerrit topic: https:/
Addressed by: https:/
Scheduler multiple workers support
Work Items
Dependency tree

* Blueprints in grey have been implemented.