New version is broken - fail to deploy with the playbook example

[danielporto]

I tried to re-run the installation playbook given as an example (and pasted below) and it failed. When I downgrade it to version 1.3.0 it runs without problems. Running the playbook on mac to deploy on Centos7.

- hosts: clients_cluster, leader, replicas
  become: no
  any_errors_fatal: true
  vars_files:
      - play-vars.yml

  roles:
    - role: Comcast.sdkman
      sdkman_dir: "{{ sdkman_folder }}"
      sdkman_auto_answer: true
      sdkman_update: true
      sdkman_install_packages:
        - { candidate: java, version: "{{ java_version }}" }
        - { candidate: ant, version: "{{ ant_version }}" }
      sdkman_flush_caches_before:
        - temp
      sdkman_flush_caches_after:
        - archives
        - broadcast
        - candidates
        - temp

EDIT: @samrocketman added code blocks to ansible example.

[elliotweiser]

Would you mind sharing some output from the playbook execution? “It failed” isn’t much to go on. While you’re at it, cranking up the verbosity would help.

[danielporto]

Sure, here is the simplified output (the -vvv one is too bloated and the error message doesnt help):

PLAY [clients_cluster, leader, replicas] **

TASK [Gathering Facts] **** ok: [capsulecorp_oc0]

TASK [Comcast.sdkman : include_tasks] ***** included: /Volumes/data/Users/dan/.ansible/roles/Comcast.sdkman/tasks/install.yml for capsulecorp_oc0

TASK [Comcast.sdkman : Determine SDKMAN vars] ***** ok: [capsulecorp_oc0]

TASK [Comcast.sdkman : Include system vars] *** ok: [capsulecorp_oc0] => (item=/Volumes/data/Users/dan/.ansible/roles/Comcast.sdkman/vars/RedHat.yml)

TASK [Comcast.sdkman : Install system packages] *** ok: [capsulecorp_oc0] => (item=curl) ok: [capsulecorp_oc0] => (item=findutils) ok: [capsulecorp_oc0] => (item=libselinux-python) ok: [capsulecorp_oc0] => (item=libstdc++) ok: [capsulecorp_oc0] => (item=tar) ok: [capsulecorp_oc0] => (item=unzip) ok: [capsulecorp_oc0] => (item=which) ok: [capsulecorp_oc0] => (item=zip)

TASK [Comcast.sdkman : include] *** skipping: [capsulecorp_oc0]

TASK [Comcast.sdkman : Check for SDKMAN installation] ***** ok: [capsulecorp_oc0]

TASK [Comcast.sdkman : Download SDKMAN] *** skipping: [capsulecorp_oc0]

TASK [Comcast.sdkman : Run SDKMAN script] ***** skipping: [capsulecorp_oc0]

TASK [Comcast.sdkman : Cleanup SDKMAN script] ***** skipping: [capsulecorp_oc0]

TASK [Comcast.sdkman : Fix permissions on SDKMAN_DIR] ***** ok: [capsulecorp_oc0]

TASK [Comcast.sdkman : include_tasks] ***** included: /Volumes/data/Users/dan/.ansible/roles/Comcast.sdkman/tasks/sdkman.yml for capsulecorp_oc0

TASK [Comcast.sdkman : Configure SDKMAN] ** ok: [capsulecorp_oc0]

TASK [Comcast.sdkman : Flush SDK caches (before)] ***** failed: [capsulecorp_oc0] (item=temp) => {"changed": false, "cmd": ". /home/goku/kaioken/sdkman/bin/sdkman-init.sh && sdk flush temp", "delta": "0:00:00.005927", "end": "2018-09-25 20:52:44.635007", "item": "temp", "msg": "non-zero return code", "rc": 127, "start": "2018-09-25 20:52:44.629080", "stderr": "find: ‘/home/goku/.sdkman/src’: No such file or directory\nfind: ‘/home/goku/.sdkman/ext’: No such file or directory\ntouch: cannot touch ‘/home/goku/.sdkman/var/delay_upgrade’: No such file or directory\n/home/goku/kaioken/sdkman/bin/sdkman-init.sh: line 103: /home/goku/.sdkman/var/candidates: No such file or directory\n/home/goku/kaioken/sdkman/bin/sdkman-init.sh: line 104: sdkman_echo_debug: command not found\n/bin/bash: sdk: command not found", "stderr_lines": ["find: ‘/home/goku/.sdkman/src’: No such file or directory", "find: ‘/home/goku/.sdkman/ext’: No such file or directory", "touch: cannot touch ‘/home/goku/.sdkman/var/delay_upgrade’: No such file or directory", "/home/goku/kaioken/sdkman/bin/sdkman-init.sh: line 103: /home/goku/.sdkman/var/candidates: No such file or directory", "/home/goku/kaioken/sdkman/bin/sdkman-init.sh: line 104: sdkman_echo_debug: command not found", "/bin/bash: sdk: command not found"], "stdout": "", "stdout_lines": []}

NO MORE HOSTS LEFT ****

PLAY RECAP **** capsulecorp_oc0 : ok=13 changed=0 unreachable=0 failed=1

[elliotweiser]

It looks like the SDKMAN_DIR variable is not being respected for some reason. It shouldn't be looking in /home/goku/.sdkman if you've overwritten sdkman_dir to /home/goku/kaioken/sdkman. I'll see if I can reproduce it.

A few other details that might be helpful:

• Python/Ansible versions
• Whether the CentOS 7 system is based on a publicly accessible docker or virtual image (if so, which one, if not, what specific version of CentOS 7)?

Thanks!

[danielporto]

local: ansible 2.6.2 python version = 2.7.15 (default, Jul 23 2018, 21:27:06) [GCC 4.2.1 Compatible Apple LLVM 9.1.0 (clang-902.0.39.2)]

remote: python: 2.7.5 centos 7 minimal installation, latest update, baremetal.

[elliotweiser]

Found the problem:

This doesn't do anything https://github.com/Comcast/ansible-sdkman/blob/master/tasks/main.yml#L7-L8

Related issue: https://github.com/ansible/ansible/issues/42990

I'll cook up a fix tomorrow.

[danielporto]

Great job!! Looking forward to seeing it in production. I've just found another bug but in v1.3.0. I want to see if it remains in v1.4.0 before submitting another issue.

[elliotweiser]

Fix is released in 1.4.1. If you find that bug you're looking for in the latest version, please don't hesitate to let us know.