Rak8s pr ansible update

[nicholasburr]

Updated to resolve deprecation warnings, and replace shell module usage with appropriate module, if available.

Updated to Kubernetes v1.17, upgrading was not allowed from 1.14 to 1.18; 1.17 minimum was required.

Updated Docker version to reflect latest available in Raspbian Buster Lite.

File: README.md Description: Updated minimum required Ansible version to 2.7.1 for the reboot module.

File: group_vars/all.yml Description:

• Updated to 1.17.0 for Kube* packages.
• Updated docker-ce version to latest included with Raspbian Buster Lite.

File: roles/cleanup/tasks/main.yml Description:

• Updated with_item to list per Ansible deprecation warning.
• Updated shell module to file module to delete pod logs.
• Updated reboot sequence to use reboot module.

File: roles/common/tasks/main.yml Description:

• Combined apt update and upgrade tasks.
• Added disabled swap, reboot to apply changes.
• Update shell module to apt module for package install.
• Updated reboot sequence to use reboot module.
• Updated when: conditional on reboot; was not working with Ansible 2.7+

File: roles/kubeadm/tasks/main.yml Description:

• Removed disable swap, it was not persisting through a reboot. Moved to common/tasks/main.yml
• Updated shell module to apt_key module.
• Updated with_item to list per Ansible deprecation warning.
• Added package locking to Kubernetes Packages. This makes the common and kubeadm roles immutable.

File: roles/master/tasks/main.yml Description:

• Added warn: false to shell command

File: upgrade.yml Description:

• Updated with_item to list per Ansible deprecation warning.
• Added warn: false to shell command
• Updated timeout due to not completing in time failures
• Updated with_item to list per Ansible deprecation warning.

A cluster was created, destroyed, created, upgraded, and destroyed locally in my network.

Using 3x rpi 3 B with Raspbian Buster Lite, release date 2020-02-13

1. ansible-playbook cluster.yml
2. ansible-playbook cleanup.yml
3. ansible-playbook cluster.yml
4. ansible-playbook upgrade.yml
5. ansible-playbook cleanup.yml

[chris-short]

I'm super excited about this PR, @nicholasburr. I wish I had time to standup my test rig this weekend but, I did yard work yesterday and I've had a beautiful day with the family here. Let me know when this is ready and I'll try to work on molecule testing.

[nicholasburr]

No worries @chris-short, this should be good for now. I am working on the dashboard, but I will submit a second PR for it.

Let me know if you need help with anything, and I will see what I can do.

[trentprynn]

Cloned @nicholasburr's PR and ran the cluster playbook on a fresh raspbian lite buster 3 node install (master and two workers). Went off without a hitch and am able to see all three nodes using kubectl. Dashboard is not working as Nicholas mentioned above.