Provisioning of nodes resets their pending_addition flag
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Invalid
|
High
|
Fuel Python (Deprecated) |
Bug Description
Can be reproduced on Fuel UI on custom #547 ISO.
Steps to reproduce:
1. create default env, add some node(s). The node(s) will have pending_addition = True attribute to determine they are not deployed.
2. start provisioning task for the environment and wait for its completion
Expected result: provisioned nodes can be deleted from the environment without deployment; user can configure disks/interfaces for provisioned node, change its roles
Actual result: provisioned node can not be deleted from the environment, just can be marked as pending deletion, that requires env deployment to delete the node; node disks/interface
tags: | added: covered-by-bp |
I'm not sure about the following statement:
> user can configure disks/interfaces for provisioned node, change its roles
We do partitioning during provisioning, so any further changes to disks will not make any affect. And since partitioning depends on roles (each node may have different volumes), ability to change node roles for provisioned nodes will lead to unexpected results.
Though I agree that we should be able to remove provisioned nodes from the env.