No description
Find a file
Paul-Henri Froidmont e7eed484e1 Merge branch 'tinc-kubeadm' into 'master'
Use tinc instead of peervpn and create cluster with kubeadm

See merge request phfroidmont/self-hosting!1
2018-09-26 19:29:46 +02:00
group_vars Remove old roles and fix idempotency issues 2018-09-26 19:24:36 +02:00
host_vars Remove old roles and fix idempotency issues 2018-09-26 19:24:36 +02:00
inventories/staging Tinc setup 2018-09-18 04:00:12 +02:00
library Deploy ingress, lego and dashboard 2018-09-26 04:40:24 +02:00
playbooks Move kubectl config into a specific playbook and fix certs path 2018-08-02 22:24:33 +02:00
roles Remove old roles and fix idempotency issues 2018-09-26 19:24:36 +02:00
scripts Deploy ingress, lego and dashboard 2018-09-26 04:40:24 +02:00
terraform Install Kubernetes 1.11 and go back to Ubuntu Xenial until Bionic is officially supported 2018-09-25 17:07:38 +02:00
.gitignore Provision with Teraform instead of Ansible 2018-09-15 01:18:57 +02:00
ansible.cfg Install Kubernetes 1.11 and go back to Ubuntu Xenial until Bionic is officially supported 2018-09-25 17:07:38 +02:00
k8s.yml Remove old roles and fix idempotency issues 2018-09-26 19:24:36 +02:00
playbook.yml Remove Plex 2018-07-28 02:13:21 +02:00
production Add S.T.B. website 2018-07-28 02:10:23 +02:00
README.md Update documentation 2018-04-25 19:00:15 +02:00
staging Split into roles and add parameters 2017-12-08 01:45:42 +01:00
Vagrantfile Add wiki 2017-12-29 03:09:04 +01:00

Self-hosting

This project maintains the entire configuration of our self-hosted services. All configuration should be done exclusively in this repo so that everything is versioned and we have a reliable and esay way to restore the production to any given state. The deployement of the configuration is done with Ansible. Everything respects the basic Ansible principle that your configuration should be idempotent. It means that that the configuration is completely independent of the current state of the server so whatever the state of the server is, the resulting state should always be the same. Because of this you shouldn't hesitate to run Ansible often to make sure that the configuration works and the server is in the expected state. If you run ansible-playbook two times in a row, the second execution should result in no changes to be made.

Deploying the configuration

The following command deploys the complete configuration.

ansible-playbook -i production playbook.yml --ask-vault-pass

For this to work, you must of course have ansible installed and have ssh access to the server(s). You will be prompted for the vault password, ask for it if you don't have it.

Deploying specific parts of the configuration

You probably don't want to deploy the entire configuration every time you make a small change. You can deploy specific roles by providing a list of tags. Checkout playbook.yml to see which tag matches a specific role. Here is an example of deploying only the wiki and the reverse proxy:

ansible-playbook -i production playbook.yml --ask-vault-pass --tags wiki,traefik