Gathering Facts / converge action stop to work after upgrade python version from 2 to 3

[camilamacedo86]

Issue Type

• Bug report

Molecule and Ansible details

ansible: 2.8.0 and 2.8.5
molecule: 2.20 and 2.22 

Molecule installation method (one of):

• pip

Ansible installation method (one of):

• pip

Detail any linters or test runners used:

Desired Behavior

After upgrade, the python version used in the solution from 2 to 3 all tests should still work as the coverage one.

Actual Behaviour

The coverage stop to work when the only change is the python version.

--> Action: 'converge'
    
    PLAY [Build Operator in Kubernetes docker container] ***************************
    
    TASK [Gathering Facts] *********************************************************
    ok: [kind-test-local]
    
    TASK [Get existing image hash] *************************************************
    ok: [kind-test-local]
    
    TASK [Build Operator Image] ****************************************************
    changed: [kind-test-local]
    
    PLAY [Converge] ****************************************************************
    
    TASK [Gathering Facts] *********************************************************
    ok: [localhost]
    
    TASK [Delete the Operator Deployment] ******************************************
    ok: [localhost]
    
    TASK [Wait 30s for Operator Deployment to terminate] ***************************
    skipping: [localhost]
    
    TASK [Create the Operator Deployment] ******************************************
    changed: [localhost]
    
    TASK [Create the ansible.example.com/v1alpha1.Memcached] ***********************
    changed: [localhost]
    
    TASK [Wait 3m for reconciliation to run] ***************************************
    FAILED - RETRYING: Wait 3m for reconciliation to run (10 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (9 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (8 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (7 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (6 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (5 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (4 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (3 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (2 retries left).
    FAILED - RETRYING: Wait 3m for reconciliation to run (1 retries left).
    fatal: [localhost]: FAILED! => {"attempts": 10, "changed": false, "resources": [{"apiVersion": "ansible.example.com/v1alpha1", "kind": "Memcached", "metadata": {"creationTimestamp": "2019-09-20T17:00:22Z", "finalizers": ["finalizer.ansible.example.com"], "generation": 1, "name": "example-memcached", "namespace": "default", "resourceVersion": "787", "selfLink": "/apis/ansible.example.com/v1alpha1/namespaces/default/memcacheds/example-memcached", "uid": "21fb17c2-dbc8-11e9-8c04-02422e706c63"}, "spec": {"size": 3}, "status": {"conditions": [{"lastTransitionTime": "2019-09-20T17:00:25Z", "message": "Running reconciliation", "reason": "Running", "status": "False", "type": "Running"}, {"ansibleResult": {"changed": 0, "completion": "2019-09-20T17:00:40.201574", "failures": 1, "ok": 0, "skipped": 0}, "lastTransitionTime": "2019-09-20T17:00:41Z", "message": "Failed to import the required Python library (openshift) on memcached-operator-f5ffb768c-8x6d6's Python /usr/bin/python. Please read module documentation and install in the appropriate location", "reason": "Failed", "status": "True", "type": "Failure"}]}}]}

IMPORTANT

• See that others face the same issue:
  Do not pass ansible tags to non-converge playbooks #2187
• To check this in the PR: Upgrade Ansible Images  operator-framework/operator-sdk#1947 (note that each change was performed and check and when the python is changed it stop to work. )
• It has been tested by myself for few days in other PRs for the same place as well.

[camilamacedo86]

c/c @fabianvf

[fabianvf]

I think we just need to set the ansible_python_interpreter inside the operator container to point to the python3 location, instead of the default /usr/bin/python. If we installed ansible with pip3, I believe merging this PR might fix it: operator-framework/operator-sdk#1828

[camilamacedo86]

We could confirm that the this issue is not caused by molecule. Ansible will not use python3 by unless it is defined to be used by ansible_python_interpreter. See its docs: https://docs.ansible.com/ansible/latest/reference_appendices/python_3_support.html#using-python-3-on-the-managed-machines-with-commands-and-playbooks

So, the issue was occurring because besides all be done with python3 it has been executed by ansible with python2.

Closing this issue now. However, would be great if molecule could show it in an error msg.