It is not possible to execute a separate task
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Released
|
Critical
|
Ihor Kalnytskyi | ||
Mitaka |
Fix Released
|
Critical
|
Ihor Kalnytskyi | ||
Newton |
Fix Released
|
Critical
|
Ihor Kalnytskyi |
Bug Description
Detailed bug description:
As per the fix for the issue (https:/
This improves redeployment cases, but breaks a possibility for a cluster operator to execute a separate task for its own needs (via direct API request or Fuel CLI command, e.g. fuel node —env 1 —node 1 —tasks globals).
As a proposal some force-like flag can added to API/CLI to allow to execute a separate task on a node(s) (and probably to the corresponding API/CLI that triggers redeployment of changes).
At the moment the issue blocks QA activities for idempotency/
Steps to reproduce:
1. Deploy a simple env, e.g. 1 controller + 1 compute
2. Execute an arbitrary task for a node, e.g.
fuel node --node 2 --tasks openstack-
Expected result: the task is executed on the given node, last_run_
Actual result: task is not executed, last_run_
Description of the environment:
MOS 9.0, build fuel-9.
tags: |
added: area-python removed: fuel-createmirror |
Changed in fuel: | |
status: | New → Confirmed |
no longer affects: | fuel/newton |
tags: | added: on-verification |
YAQL conditions should be evaluated with empty old context for separate task runs.