Conditional role

Archives Ansible Project
1

I’m trying to modify this Ansible project and learn as I go. The role shown below is throwing errors and I’d like to clean those up, as well as make it optional. Everything in this role is concerned with installing Python and running tests, something that I don’t care about right now because I’m deploying on VM’s.

`

  • name: Bootstrap coreos hosts
    hosts: all
    gather_facts: false
    roles:
  • role: bootstrap/ansible-bootstrap
    tags: [ ansible-bootstrap ]

`

But I can’t get anything to work. If I comment it out, I get errors. If I put an override on every test, causing it to return 0 tests, I get errors. I don’t know what to do.

2

I meant to say “tasks”, not “tests”. The role file is just a list of tasks.

3

You need to tell us what errors you are getting. Ideally, post the role code as well.

“Doctor, it hurts!” is not enough to diagnose you :slight_smile:

Regards, K.

4

Hi Karl, sure thing.

So I think I have the conditional working for the role, but not I’m stuck on testing that a directory exists. Here is the role.

`

filename: roles/bootstrap/ansible-bootstrap/tasks/configure.yml

  • block:
  • name: Check if Python is installed
    raw: “[ -e {{ ansible_python_interpreter }} ] && echo ‘yes’ || echo ‘no’”
    register: python_install
    when: false

changed_when: false

  • name: Check if install tar file exists
    raw: “[ -e /tmp/pypy-{{ pypy_version }}.tar.bz2 ] && echo ‘/tmp/pypy-{{ pypy_version }}.tar.bz2’ || echo ""”
    register: pypy_tar_file
    when: false

  • name: Check if pypy directory exists
    raw: “[ -e {{ pypy_dir }} ] && echo ‘{{ pypy_dir }}’ || echo ""”
    register: pypy_directory
    when: false

  • name: Check if libtinfo is simlinked
    raw: “[ -e {{ pypy_dir }}/lib/libtinfo.so.5 ] && echo ‘{{ pypy_dir }}/lib/libtinfo.so.5’ || echo ""”
    register: libtinfo_symlink
    when: false
    ignore_errors: yes

  • name: Download PyPy
    raw: wget -O /tmp/pypy-{{ pypy_version }}.tar.bz2 https://bitbucket.org/pypy/pypy/downloads/pypy-{{ pypy_version }}-linux64.tar.bz2
    when: pypy_tar_file.stdout == “”

  • name: Extract PyPy
    raw: mkdir -p {{ pypy_dir }} && tar xf /tmp/pypy-{{ pypy_version }}.tar.bz2 -C {{ pypy_dir }}/ --strip-components=1
    args:
    warn: false
    when: pypy_directory.stdout == “”

  • name: Symlink libncurses as libtinfo.so.5
    raw: mkdir -p {{ pypy_dir }}/lib && ln -s /lib64/libncurses.so.5.9 {{ pypy_dir }}/lib/libtinfo.so.5
    args:
    warn: false
    when: libtinfo_symlink.stdout == “”

  • name: Install PyPy
    script: bootstrap.sh {{ pypy_dir }} {{ ansible_python_interpreter }}
    when: python_install.stdout == ‘yes’

`

The error message seems to implicate the ansible_python_interpreter variable, since it has the path /home/core/bin/python.

I’ve also tried this popular variation:

`

  • name: Check if Python is installed
    stat:
    path: ansible_python_interpreter
    register: python_install
    when: false

`

And updated the when at the bottom. Also tried using - debug.

`

when: file_details.stat.exists

`

No matter what I do, the output is always the same, which leads me to conclude the error must be somewhere else, but I can’t see where else it could be coming from. Is it because everything above it was skipped?

`

Bringing machine ‘etcd-01’ up with ‘virtualbox’ provider…

Bringing machine ‘kube-master-01’ up with ‘virtualbox’ provider…

Bringing machine ‘kube-worker-01’ up with ‘virtualbox’ provider…

Bringing machine ‘kube-worker-02’ up with ‘virtualbox’ provider…

==> etcd-01: Checking if box ‘coreos-stable’ is up to date…

==> etcd-01: [vagrant-hostmanager:guests] Updating hosts file on active guest virtual machines…

==> etcd-01: Machine already provisioned. Run vagrant provision or use the --provision

==> etcd-01: flag to force provisioning. Provisioners marked to run always will still run.

==> kube-master-01: Checking if box ‘coreos-stable’ is up to date…

==> kube-master-01: [vagrant-hostmanager:guests] Updating hosts file on active guest virtual machines…

==> kube-master-01: Machine already provisioned. Run vagrant provision or use the --provision

==> kube-master-01: flag to force provisioning. Provisioners marked to run always will still run.

==> kube-worker-01: Checking if box ‘coreos-stable’ is up to date…

==> kube-worker-01: [vagrant-hostmanager:guests] Updating hosts file on active guest virtual machines…

==> kube-worker-01: Machine already provisioned. Run vagrant provision or use the --provision

==> kube-worker-01: flag to force provisioning. Provisioners marked to run always will still run.

==> kube-worker-02: Checking if box ‘coreos-stable’ is up to date…

==> kube-worker-02: [vagrant-hostmanager:guests] Updating hosts file on active guest virtual machines…

==> kube-worker-02: Machine already provisioned. Run vagrant provision or use the --provision

==> kube-worker-02: flag to force provisioning. Provisioners marked to run always will still run.

PLAY [Bootstrap coreos hosts] **********************************************************************************************************************************************************

TASK [bootstrap/ansible-bootstrap : Check if Python is installed] **********************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Check if install tar file exists] ******************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Check if pypy directory exists] ********************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Check if libtinfo is simlinked] ********************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Download PyPy] *************************************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Extract PyPy] **************************************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Symlink libncurses as libtinfo.so.5] ***************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Install PyPy] **************************************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Test if Python is installed] ***********************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

TASK [bootstrap/ansible-bootstrap : Gather ansible facts] ******************************************************************************************************************************

skipping: [kube-master-01]

skipping: [etcd-01]

skipping: [kube-worker-01]

skipping: [kube-worker-02]

PLAY [Create certificates for Kubernetes componentes] **********************************************************************************************************************************

TASK [Gathering Facts] *****************************************************************************************************************************************************************

fatal: [etcd-01]: FAILED! => {“changed”: false, “module_stderr”: “Connection to 127.0.0.1 closed.\r\n”, “module_stdout”: “/bin/sh: /home/core/bin/python: No such file or directory\r\n”, “msg”: “MODULE FAILURE”, “rc”: 127}

fatal: [kube-worker-01]: FAILED! => {“changed”: false, “module_stderr”: “Connection to 127.0.0.1 closed.\r\n”, “module_stdout”: “/bin/sh: /home/core/bin/python: No such file or directory\r\n”, “msg”: “MODULE FAILURE”, “rc”: 127}

fatal: [kube-master-01]: FAILED! => {“changed”: false, “module_stderr”: “Connection to 127.0.0.1 closed.\r\n”, “module_stdout”: “/bin/sh: /home/core/bin/python: No such file or directory\r\n”, “msg”: “MODULE FAILURE”, “rc”: 127}

fatal: [kube-worker-02]: FAILED! => {“changed”: false, “module_stderr”: “Connection to 127.0.0.1 closed.\r\n”, “module_stdout”: “/bin/sh: /home/core/bin/python: No such file or directory\r\n”, “msg”: “MODULE FAILURE”, “rc”: 127}

PLAY RECAP *****************************************************************************************************************************************************************************

etcd-01 : ok=0 changed=0 unreachable=0 failed=1

kube-master-01 : ok=0 changed=0 unreachable=0 failed=1

kube-worker-01 : ok=0 changed=0 unreachable=0 failed=1

kube-worker-02 : ok=0 changed=0 unreachable=0 failed=1

`

This is the repo I’ve been working from: GitHub - sebiwi/kubernetes-coreos: Automatic local Kubernetes cluster provisioning and configuration on CoreOS using Ansible and Vagrant

Starts with the Makefile, then enters kubernetes.yml, and the subject of concern is the very first role listed.