Deployment Plan for Home Cloud

1. Goal

The goal is to set up a personal cloud environment on your two hosts, Mulder and Scully. This involves deploying Gitea (a self-hosted Git service) on Mulder, and Keycloak (an identity and access management solution) on Scully. All services should be accessible via HTTPS with Let's Encrypt certificates and subdomain-based routing. We will also set up a WireGuard VPN with a web interface for secure access to your network.

2. Phased Deployment Plan

Phase 1: Network Infrastructure on Scully (Completed)

Goal: Deploy the common and network roles on Scully. The network role will set up Nginx Proxy Manager (for HTTPS and subdomain routing) and WireGuard Easy (for VPN with web UI).
Host and Role Assignments:
- Scully: common, podman, network
Configuration Files:
- inventory/hosts.yml: Defines Mulder and Scully, their connection details, and role-specific variables.
- playbooks/main.yml: Modified to execute the common, podman, and network roles on Scully.
- group_vars/all.yml: Contains common variables like the domain name and service credentials.
Execution Plan:
1. Run the playbook to deploy the common, podman, and network roles on Scully. (Podman and Portainer are now successfully installed).
2. After successful execution, verify the network services.

Phase 2: Gitea and Keycloak Deployment (Next)

Goal: Deploy Gitea on Mulder and Keycloak on Scully, along with their respective PostgreSQL databases.
Host and Role Assignments:
- Mulder: common, podman, postgres, gitea
- Scully: common, podman, postgres, keycloak (in addition to network)
Dependencies: This phase depends on the successful completion of Phase 1 and the availability of the domain name.
Next Steps: Once Phase 1 is complete, we will update the playbooks/main.yml and group_vars/all.yml to include the postgres, gitea, and keycloak roles.

3. What We Still Need

Your Domain Name: Please provide the domain name you want to use for your personal cloud (e.g., my-cloud.com). (Already provided as ai-eifel.de).

4. Dry-Run

Dry runs will be performed where appropriate, but direct execution will be used for tasks that require actual changes to the system.

2.3 KiB Raw Blame History