Zun

Container resize needs resource claiming in compute node

Registered by hongbin

Currently, we allow users to update the 'cpu' and 'memory' of the container but we don't check if the host has enough resources for this update. As a result, the container might take more resources than available.

We should perform resource checking while updating the container (similar as when we create the container), each update request should go through resource_tracker and do resource claiming (https://github.com/openstack/zun/blob/master/zun/compute/claims.py) in the compute node, which ensure the host has enough resources.

NOTE: This is how we do resource claiming when creating container: https://review.openstack.org/#/c/463638/11/zun/compute/manager.py

Blueprint information

Status:
Complete
Approver:
hongbin
Priority:
Essential
Drafter:
hongbin
Direction:
Approved
Assignee:
feng.shengqin@zte.com.cn
Definition:
New
Series goal:
Accepted for rocky
Implementation:
Implemented
Milestone target:
None
Started by
hongbin
Completed by
feng.shengqin@zte.com.cn

Related branches

Sprints

Whiteboard

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

Addressed by: https://review.openstack.org/531651
    [WIP] claim resources when updating cpu/memory

(?)

Work Items

This blueprint contains Public information 
Everyone can see this information.

Subscribers

No subscribers.