Iac

Status: in progress.

A resilient remote-access gateway built so that no single component — web tier, application tier, or database — is a single point of failure:

• Multiple Guacamole nodes behind an Nginx load balancer, so a node failure doesn’t take down access
• MariaDB Galera cluster as the authentication/connection database backend — synchronous multi-master replication, no single DB instance to lose
• Session recording for connection auditing and compliance
• Infrastructure as Code — the whole stack defined and reproducible, not hand-built

The interesting part of this project isn’t installing Guacamole — that’s well documented. It’s getting the Galera cluster quorum right (odd node count, proper arbitrator handling) and making sure failover at every layer — Nginx, Guacamole, and the database — actually works under a node failure, not just in theory.

The problem

There’s something innate in some people — the drive to fix whatever is annoying. Annoying to some, but apparently default behavior for others.

Since I started my journey with Proxmox, I was amazed by its capabilities, and every release improved on that foundation. But there was trouble in this newfound paradise: one of the most repetitive tasks was creating a VM almost identical to the previous one. Same base OS, same baseline packages, same manual clicking through the same wizard.