Split nova role into nova control and nova compute
In current playbook, nova control services and nova compute are deployed together. However, in reality they could have difference configuration parameters. For example, compute_driver is not necessary for nova_api. In addition, in a cloud if we want to different nova-compute, the current playbook will re-deploy the nova control container as well. Therefore, it is better to have separate roles for nova control (inducing nova_api, nova_conductor, ..) and nova-compute
Blueprint information
- Status:
- Not started
- Approver:
- None
- Priority:
- Undefined
- Drafter:
- Hui Kang
- Direction:
- Needs approval
- Assignee:
- None
- Definition:
- New
- Series goal:
- None
- Implementation:
-
Unknown
- Milestone target:
- None
- Started by
- Completed by
Related branches
Related bugs
Sprints
Whiteboard
Another use case is that: say I have already deployed an openstack whose control plane nodes are centos binary. Later I would like to add some compute nodes whose container OS are ubuntu source. If I use the existing playbook, there are two potential options (1) add variable definition in the inventory file. Although it works, this is kind of tedious. (2) Run playbook twice, each time with different inventory file (e.g., one includes compute nodes, the other does not). This way the playbook may fail because either (1) tasks on compute nodes fails due to no control node in the playbook (2) the control node will be restarted with the new container type. - Hui
