At least one controller is required to redeploy environment
Bug #1410316 reported by
Julia Aranovich
This bug affects 2 people
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Invalid
|
Medium
|
Fuel UI Team | ||
Mitaka |
Invalid
|
Medium
|
Fuel UI Team |
Bug Description
Deploy multi-node env, then mark controller as 'pending deletion'. Add new controller and deploy environment. There will be an error (see an attached screenshot): "Not enough deployed controllers - deployed cluster requires at least 1 deployed controller."
From my point of view this message should appear in Deploy Changes dialog and block redeployment (lock Deploy button) as we do for 'New' environments.
Changed in fuel: | |
status: | New → Confirmed |
Changed in fuel: | |
importance: | Medium → Low |
tags: | added: qa-agree-7.0 |
tags: | added: release-notes |
Changed in fuel: | |
status: | Confirmed → Won't Fix |
Changed in fuel: | |
milestone: | 7.0 → 8.0 |
status: | Won't Fix → Confirmed |
no longer affects: | fuel/8.0.x |
tags: | removed: release-notes |
tags: | added: area-ui |
summary: |
- [ui] At least one controller is required to redeploy environment + At least one controller is required to redeploy environment |
Changed in fuel: | |
status: | Confirmed → Won't Fix |
To post a comment you must log in.
The above messages are moved to 'limits' object in this review request: https:/ /review. openstack. org/#/c/ 126930/
Unfortunately the situation you describe here is not covered by that request. Let me explain: all limits are checked assuming deployment is successful. I.e. for controller we have min: 1. So in new cluster we add new controller and so number of controllers with pending_addition is 1. The limit checker is happy because this is correct -- after successful deployment there will be 1 controller.
What you describe is to check 'worst case' limits, i.e. that for some time during deployment there will be no controller at all. But when deploying new cluster there is also such situation: for some time there is no controller at all but the deployment procedure is correct!
The solution I see is to modify limit checking code so that it counts number of nodes differently for fresh cluster and for a redeployed one. If you have other suggestions then please share :)