Ansible fails to locate templates in role's /templates directory [duplicate] - ansible

Question

Does ansible pass Role Default variables to the Handlers within the same Role?
Here's a minimal excerpt of the playbook that has the issue:
Role hierarchy
- playbook.yml
- roles/
- gunicorn/
- defaults/
- main.yml
- handlers/
- main.yml
- code-checkout/
- tasks/
- main.yml
Here's the file contents
gunicorn/defaults/main.yml
---
gu_log: "/tmp/gunicorn.log"
gunicorn/handlers/main.yml
---
- name: Clear Gunicorn Log
shell: rm {{ gu_log }}
finalize/tasks/main.yml
---
- name: Test Handlers
shell: ls
notify:
- Restart Gunicorn
playbook.yml
---
- name: Deploy
hosts: webservers
tasks:
- include: roles/finalize/tasks/main.yml
handlers:
- include: roles/gunicorn/handlers/main.yml
AFAIK everything looks good. However, I get this error during the playbook execution
FAILED! => {"failed": true, "msg": "the field 'args' has an invalid
value, which appears to include a variable that is undefined. The
error was: 'gu_log' is undefined\n\nThe error appears to have been in
'/roles/gunicorn/handlers/main.yml':
line 3, column 3, but may\nbe elsewhere in the file depending on the
exact syntax problem.\n\nThe offending line appears to be:\n\n\n-
name: Restart Gunicorn\n ^ here\n"}
Using Ansible 2.2 on Ubuntu 12.04 LTS
Here's a modified version of techraf's script that creates all the directories and demonstrates my issue
#!/bin/bash
mkdir -p ./rtindru-test/roles/gunicorn
mkdir -p ./rtindru-test/roles/gunicorn/defaults
mkdir -p ./rtindru-test/roles/gunicorn/handlers
mkdir -p ./rtindru-test/roles/finalize/tasks
cat >./rtindru-test/roles/finalize/tasks/main.yml <<HANDLERS_END
---
- name: Test Handlers
shell: rm {{ gu_log }}
HANDLERS_END
cat >./rtindru-test/roles/gunicorn/handlers/main.yml <<HANDLERS_END
---
- name: Clear Gunicorn Log
shell: rm {{ gu_log }}
HANDLERS_END
cat >./rtindru-test/roles/gunicorn/defaults/main.yml <<DEFAULTS_END
---
gu_log: "/tmp/gunicorn.log"
DEFAULTS_END
cat >./rtindru-test/playbook.yml <<PLAYBOOK_END
---
- name: Deploy
hosts: localhost
tasks:
- include: roles/finalize/tasks/main.yml
handlers:
- include: roles/gunicorn/handlers/main.yml
PLAYBOOK_END
touch /tmp/gunicorn.log
ls -l /tmp/gunicorn.log
ansible-playbook ./rtindru-test/playbook.yml
ls -l /tmp/gunicorn.log
Output
PLAY [Deploy]
TASK [setup]
******************************************************************* ok: [localhost]
TASK [Test Handlers]
*********************************************************** fatal: [localhost]: FAILED! => {"failed": true, "msg": "the field 'args' has
an invalid value, which appears to include a variable that is
undefined. The error was: 'gu_log' is undefined\n\nThe error appears
to have been in '/rtindru-test/roles/finalize/tasks/main.yml': line 2,
column 3, but may\nbe elsewhere in the file depending on the exact
syntax problem.\n\nThe offending line appears to be:\n\n---\n- name:
Test Handlers\n ^ here\n"} to retry, use: --limit
#/rtindru-test/playbook.retry
PLAY RECAP
********************************************************************* localhost : ok=1 changed=0 unreachable=0
failed=1

Answer 1

You are neither defining nor using any roles. With the following task:
- include: roles/finalize/tasks/main.yml
you are only including a tasks file into your playbook. It has nothing to do with roles.
To assign a role you should specify a list of roles for a play (one or more):
role:
- my_role1
- my_role2
Please have a look at the documentation on roles and feel free to use the playbook and structure as created by the below script.
Does ansible pass Role Default variables to the Handlers within the same Role?
Yes it does.
For a proof run the following bash script which creates and runs a minimal example. It takes the contents of gunicorn/defaults/main.yml and gunicorn/handlers/main.yml from the question intact and adds missing components: the tasks and the playbook. It creates a file to be removed and runs the playbook.
#!/bin/bash
mkdir -p ./so41285033/roles/gunicorn
mkdir -p ./so41285033/roles/gunicorn/defaults
mkdir -p ./so41285033/roles/gunicorn/handlers
mkdir -p ./so41285033/roles/gunicorn/tasks
cat >./so41285033/roles/gunicorn/tasks/main.yml <<TASKS_END
---
- debug:
changed_when: true
notify: Clear Gunicorn Log
TASKS_END
cat >./so41285033/roles/gunicorn/handlers/main.yml <<HANDLERS_END
---
- name: Clear Gunicorn Log
shell: rm {{ gu_log }}
when: "'apiservers' not in group_names"
HANDLERS_END
cat >./so41285033/roles/gunicorn/defaults/main.yml <<DEFAULTS_END
---
gu_log: "/tmp/gunicorn.log"
DEFAULTS_END
cat >./so41285033/playbook.yml <<PLAYBOOK_END
---
- hosts: localhost
gather_facts: no
connection: local
roles:
- gunicorn
PLAYBOOK_END
touch /tmp/gunicorn.log
ls -l /tmp/gunicorn.log
ansible-playbook ./so41285033/playbook.yml
ls -l /tmp/gunicorn.log
The result:
-rw-r--r-- 1 techraf wheel 0 Dec 23 07:57 /tmp/gunicorn.log
[WARNING]: Host file not found: /etc/ansible/hosts
[WARNING]: provided hosts list is empty, only localhost is available
PLAY [localhost] ***************************************************************
TASK [gunicorn : debug] ********************************************************
ok: [localhost] => {
"msg": "Hello world!"
}
RUNNING HANDLER [gunicorn : Clear Gunicorn Log] ********************************
changed: [localhost]
[WARNING]: Consider using file module with state=absent rather than running rm
PLAY RECAP *********************************************************************
localhost : ok=2 changed=2 unreachable=0 failed=0
ls: /tmp/gunicorn.log: No such file or directory
Interpretation:
Before running the playbook the file /tmp/gunicorn.log was created and its existence verified:
-rw-r--r-- 1 techraf wheel 0 Dec 23 07:57 /tmp/gunicorn.log
After running the playbook the file /tmp/gunicorn.log does not exist:
ls: /tmp/gunicorn.log: No such file or directory
Ansible correctly passed the variable gu_log value to the Clear Gunicorn Log handler which removed the file.
Final remark:
The problem described in question is impossible to reproduce, because the question does not contain complete nor verifiable example in the meaning of MCVE.