[k8s] Failed to find exact match for v1.DeploymentConfig by [kind, name, singularName, shortNames]

[amadeus-chuot]

SUMMARY

I'm not able to use k8s to create a DeploymentConfig on Openshift 3.6.

Error messages are:

• ResourceNotFoundError('No matches found for {}'.format(kwargs))
• Failed to find exact match for v1.DeploymentConfig by [kind, name, singularName, shortNames]

Please help to fix or advise if it is not supported.

The same setup works when creating Pods or Secrets.

ISSUE TYPE

• Bug Report

COMPONENT NAME

k8s

ANSIBLE VERSION

ansible 2.7.10
  config file = None
  configured module search path = ['/Users/chuot/.ansible/plugins/modules', '/usr/share/ansible/plugins/modules']
  ansible python module location = /Users/chuot/Dev/Ansible/ansible-virtualenv/lib/python3.6/site-packages/ansible
  executable location = /Users/chuot/Dev/Ansible/ansible-virtualenv/bin/ansible
  python version = 3.6.8 (default, Apr 11 2019, 22:58:09) [GCC 4.2.1 Compatible Clang 7.0.1 (tags/RELEASE_701/final)]

CONFIGURATION

(empty)

OS / ENVIRONMENT

MacOS 10.13.6

$ uname -a
Darwin NCELRND0546 17.7.0 Darwin Kernel Version 17.7.0: Thu Dec 20 21:47:19 PST 2018; root:xnu-4570.71.22~1/RELEASE_X86_64 x86_64 i386 MacBookPro14,1 Darwin

Gentoo Prefix

Virtualenv for ansible

STEPS TO REPRODUCE

Here is a small test config:

playbook.yml

---
- name: Test case for creating a DeploymentConfig with Ansible
  hosts: all
  gather_facts: no
  vars:
    k8s_url: https://172.30.19.49:8443/
    k8s_ns: myproject
    k8s_apikey: xxx
  tasks:
    - name: test DeploymentConfig
      delegate_to: 127.0.0.1
      run_once: true
      k8s:
        state: present
        resource_definition: "{{ lookup('template', 'test.yml') }}"
        host: "{{ k8s_url }}"
        namespace: "{{ k8s_ns }}"
        api_key: "{{ k8s_apikey }}"
        verify_ssl: false
...

test.yml:

kind: DeploymentConfig
apiVersion: v1
metadata:
  name: totoDc
spec:
  template:
    metadata:
      labels:
        name: totoTemplate
    spec:
      terminationGracePeriodSeconds: 120
      containers:
      - name: totoContainer
        image: my/image:thetag
        imagePullPolicy: IfNotPresent
  replicas: 1

and a minimalist inventory which is actually not used (since running on control node):
inventory:

[all]
toto

EXPECTED RESULTS

The DeploymentConfig should be created

ACTUAL RESULTS

$ ansible-playbook -i inventory -vvvv playbook.yml
ansible-playbook 2.7.10
  config file = None
  configured module search path = ['/Users/chuot/.ansible/plugins/modules', '/usr/share/ansible/plugins/modules']
  ansible python module location = /Users/chuot/Dev/Ansible/ansible-virtualenv/lib/python3.6/site-packages/ansible
  executable location = /Users/chuot/Dev/Ansible/ansible-virtualenv/bin/ansible-playbook
  python version = 3.6.8 (default, Apr 11 2019, 22:58:09) [GCC 4.2.1 Compatible Clang 7.0.1 (tags/RELEASE_701/final)]
No config file found; using defaults
setting up inventory plugins
/Users/chuot/Dev/Ansible/oracle_observers/molecule/default/inventory did not meet host_list requirements, check plugin documentation if this is unexpected
/Users/chuot/Dev/Ansible/oracle_observers/molecule/default/inventory did not meet script requirements, check plugin documentation if this is unexpected
Parsed /Users/chuot/Dev/Ansible/oracle_observers/molecule/default/inventory inventory source with ini plugin
Loading callback plugin default of type stdout, v2.0 from /Users/chuot/Dev/Ansible/ansible-virtualenv/lib/python3.6/site-packages/ansible/plugins/callback/default.py

PLAYBOOK: playbook.yml **************************************************************************************************************************************************************
1 plays in playbook.yml

PLAY [Test case for creating a DeploymentConfig with Ansible] ***********************************************************************************************************************
META: ran handlers

TASK [test DeploymentConfig] ********************************************************************************************************************************************************
task path: /Users/chuot/Dev/Ansible/oracle_observers/molecule/default/playbook.yml:10
File lookup using /Users/chuot/Dev/Ansible/oracle_observers/templates/test.yml as file
<127.0.0.1> ESTABLISH LOCAL CONNECTION FOR USER: chuot
<127.0.0.1> EXEC /bin/sh -c 'echo ~chuot && sleep 0'
<127.0.0.1> EXEC /bin/sh -c '( umask 77 && mkdir -p "` echo /Users/chuot/.ansible/tmp/ansible-tmp-1555088971.376927-162544634893078 `" && echo ansible-tmp-1555088971.376927-162544634893078="` echo /Users/chuot/.ansible/tmp/ansible-tmp-1555088971.376927-162544634893078 `" ) && sleep 0'
Using module file /Users/chuot/Dev/Ansible/ansible-virtualenv/lib/python3.6/site-packages/ansible/modules/clustering/k8s/k8s.py
<127.0.0.1> PUT /Users/chuot/.ansible/tmp/ansible-local-33226s8wbs9mh/tmph00zv9rk TO /Users/chuot/.ansible/tmp/ansible-tmp-1555088971.376927-162544634893078/AnsiballZ_k8s.py
<127.0.0.1> EXEC /bin/sh -c 'chmod u+x /Users/chuot/.ansible/tmp/ansible-tmp-1555088971.376927-162544634893078/ /Users/chuot/.ansible/tmp/ansible-tmp-1555088971.376927-162544634893078/AnsiballZ_k8s.py && sleep 0'
<127.0.0.1> EXEC /bin/sh -c '/Users/chuot/Dev/Ansible/ansible-virtualenv/bin/python /Users/chuot/.ansible/tmp/ansible-tmp-1555088971.376927-162544634893078/AnsiballZ_k8s.py && sleep 0'
<127.0.0.1> EXEC /bin/sh -c 'rm -f -r /Users/chuot/.ansible/tmp/ansible-tmp-1555088971.376927-162544634893078/ > /dev/null 2>&1 && sleep 0'
The full traceback is:
  File "/var/folders/lq/mtrm60rd3xb75c4dv54rppk00000gp/T/ansible_k8s_payload_bchndbu9/ansible_k8s_payload.zip/ansible/module_utils/k8s/common.py", line 194, in find_resource
    return self.client.resources.get(api_version=api_version, short_names=[kind])
  File "/Users/chuot/Dev/Ansible/ansible-virtualenv/lib/python3.6/site-packages/openshift/dynamic/client.py", line 637, in get
    raise ResourceNotFoundError('No matches found for {}'.format(kwargs))

fatal: [toto -> 127.0.0.1]: FAILED! => {
    "changed": false,
    "invocation": {
        "module_args": {
            "api_key": "VALUE_SPECIFIED_IN_NO_LOG_PARAMETER",
            "cert_file": null,
            "context": null,
            "force": false,
            "host": "https://172.30.19.49:8443/",
            "key_file": null,
            "kubeconfig": null,
            "merge_type": null,
            "password": null,
            "ssl_ca_cert": null,
            "state": "present",
            "username": null,
            "verify_ssl": false
        }
    },
    "msg": "Failed to find exact match for v1.DeploymentConfig by [kind, name, singularName, shortNames]"
}

NO MORE HOSTS LEFT ******************************************************************************************************************************************************************
	to retry, use: --limit @/Users/chuot/Dev/Ansible/oracle_observers/molecule/default/playbook.retry

PLAY RECAP **************************************************************************************************************************************************************************
toto                       : ok=0    changed=0    unreachable=0    failed=1

[ansibot]

Files identified in the description:

• lib/ansible/modules/clustering/k8s/k8s.py

If these files are inaccurate, please update the component name section of the description or use the !component bot command.

click here for bot help

[ansibot]

cc @chouseknecht @fabianvf @flaper87 @maxamillion @willthames
click here for bot help

[willthames]

Can you add a task before that one with

- python_requirements_facts:
    dependencies:
      - openshift
      - kubernetes

I get a different response from openshift 3.6 (using openshift/origin:v3.6.1 container) so would be good to be able to reproduce your exact issue

[willthames]

Ah, reproduced it by upgrading my openshift to 0.8.7

[willthames]

So, I've got to the bottom of this but I don't know what to do about it.

Basically in the /oapi/v1 response, both generatedeploymentconfig and deploymentconfig resources are of kind DeploymentConfig. When looking for what resource to use for kind DeploymentConfig, multiple resources are returned so a ResourceNotUniqueError error is returned (you can see this in the stack trace you provide).

generatedeploymentconfig is deprecated in openshift 3.7, 3.6 is an incredibly old openshift version, so I'm not sure what to do from here - @fabianvf might be able to help in the openshift/openshift-restclient-python repository

[willthames]

I can confirm this issue goes away running against openshift/origin:v3.7

It may still be worth addressing in case of future API changes, although I'd hope that the mappings are more consistent these days.

[amadeus-chuot]

Thanks for confirming the issue and that it works with 3.7.
Just to be sure, you talk about ResourceNotUniqueError but my stack shows a ResourceNotFoundError?

I need to deploy in production on 3.6, before the upgrade to 3.7.
Can I tweak the namespace to hide this generatedeploymentconfig? Or do you think about any other workaround?

[amadeus-chuot]

Also, since I'm finding nothing about this generatedeploymentconfig, I suppose nobody is using it and would it be possible to filter it out from the Ansible module?

[willthames]

This problem should really be addressed upstream in the openshift library - we don't have enough control to filter this out.

If you're seeing ResourceNotFoundError then you may need a newer version of the openshift library.

[amadeus-chuot]

Ok this is solved.

After digging into the openshift lib code, I found that search() was returning 3 objects:
[<Resource(v1/deploymentconfigs)>, <Resource(v1/generatedeploymentconfigs)>, <Resource(apps.openshift.io/v1/deploymentconfigs)>]

I managed to fix my issue by using apiVersion: apps.openshift.io/v1 instead of v1

Thanks for the help

[fabianvf]

@amadeus-chuot thanks a ton for your additional investigation! I'll create a follow-up issue to surface these errors a little better, you shouldn't have to dig around manually to find that information.

[nhenderson]

For anyone interested, here's a non-ideal work-around for my similar issue:

Failed to find exact match for v1.Template by [kind, name, singularName, shortNames]

It should work for the v1.DeploymentConfig issue as well:

- name: Create resources from OpenShift template
  shell: |
    oc create -n {{ openshift_project_name }} -f - <<-'EOF'
    {{ lookup('template', role_path + '/templates/main.yaml.j2') }}
    EOF