Tailmox 1.2.0: A Cleaner Way to Connect Proxmox Clusters Across the Globe

If you've ever tried stretching your Proxmox cluster across cities, countries, or even just different ISPs, you know the pain. WAN latency, VPN configuration, NAT traversal, awkward ports—all of it turns what should be a neat homelab setup into a weekend-long troubleshooting marathon. But that's exactly where Tailmox 1.2.0 enters the conversation. Tailmox is a community-driven tool that connects Proxmox hosts through Tailscale, a zero-config VPN built on WireGuard. The latest release, version 1.2.0, changes the game by shifting its architecture to use tailscale serve, allowing users to bypass some of Proxmox's native networking limitations. The result? You can now access the Proxmox web interface without needing to tack on port 8006 to the URL or wrestle with certificates tied to pveproxy. It's not just cleaner—it's smarter. ## So what exactly is Tailmox doing here? At its core, Tailmox simplifies secure remote access to Proxmox hosts by integrating with Tailscale. Before this release, Tailmox relied more directly on modifying the Proxmox stack (specifically pveproxy) to serve the web UI over Tailscale. Now with tailscale serve, the script decouples those services, removing the need for deep integration hacks or persistent proxy reconfigurations. That might sound like a small tweak, but for anyone who's had to manage distributed nodes across geographic regions, it's a big deal. Tailmox effectively turns any Proxmox node—even one behind CGNAT or in a remote location—into a member of a virtual mesh network. You can cluster, replicate, migrate, and manage your nodes from anywhere, as long as there's a basic internet connection. ## Global clustering: great on paper, tricky in practice In the homelab world, clustering Proxmox nodes across different geographies is the stuff of legends and headaches. Willjasen, the creator behind Tailmox, dropped a solid reminder in the release thread: yes, it's possible—but think before you dive in. Tailmox now includes a dedicated README section that outlines considerations for global clustering. It's not trying to be everything to everyone. The tool is ideal for setups where you're not relying on full-blown HA (High Availability) or Ceph-backed storage that needs extremely low latency. Instead, Tailmox fits scenarios where VM replication and migrations happen on a slower cadence—or where simplified backup flows are the priority. Think of someone moving a 20TB VM from Europe to the US. Tailmox helped facilitate that by seeding the data with ZFS replication, then using the Proxmox tools to finalize the migration with just a few minutes of downtime. That kind of move is nearly impossible to do safely without setting up a lot of infrastructure—or trusting your ISP to not be a bottleneck. ## Not ready for prod—yet It's worth pointing out that even its creator doesn't currently use Tailmox in a production environment. In fact, willjasen stresses that the script is best run on a fresh Proxmox install, before any heavy customization. That's not a red flag—it's just honest. Tailmox is still a work in progress, and using it on a host you can't afford to lose might be a bad call. Instead, the recommended approach is "Proxmox-in-Proxmox." Yep, it's exactly what it sounds like. Willjasen tests Tailmox by spinning up virtual Proxmox VMs inside a Proxmox cluster. He snapshots them, runs the script, and reverts to the snapshot as needed to iterate fast. This kind of meta-homelab setup makes it easier to trial Tailmox without wrecking your real nodes—and it's a great tip for anyone experimenting with similar tools. ## Why not just use qm remote-migrate? Some folks in the community rightly pointed out that Proxmox already supports migration tools like qm remote-migrate. Why reinvent the wheel? The short answer: size, flexibility, and security. The qm command is fine for smaller VMs and basic migrations. But when you're dealing with large volumes—or when you want to stage the data first using ZFS replication—the vanilla tools don't quite cut it. Plus, Tailmox wraps this functionality in a much cleaner, more secure way using Tailscale, which avoids the pain of punching firewall holes and exposes zero ports to the public internet. There's also the matter of persistence. Tailmox can be a way to futureproof your setup. Want to move your node to another building, city, or even continent? As long as it can connect to Tailscale, it can still show up as part of the cluster. ## What about overhead and performance? Some users raised the question of virtualization overhead when running Proxmox inside Proxmox, especially for test setups. But according to Willjasen, it hasn't been an issue. The "nested" VMs work well enough for functional testing of clustering, networking, and script logic. Of course, this isn't meant for benchmarking production workloads—but for testing Tailmox? It's perfect. As for the real-world usage on low-power devices, one commenter mentioned running Tailmox on a thin client PC to act as a remote node. The plan? Connect it to a virtual cluster back home and use it as a real-world testbed. That's exactly the kind of flexible, modular thinking Tailmox encourages. ## Tailmox isn't trying to be Hyper-V—or Ceph Some folks compared the setup to Hyper-V's failover clustering or Proxmox's own HA stack using Ceph. But that's not what Tailmox is built for. Instead, think of it as a glue layer. It helps Proxmox hosts talk to each other securely, no matter where they're located. And while it doesn't replace proper HA or shared storage, it does offer a lightweight path to VM replication, remote backups, and migration scenarios that previously required a lot more sweat and swearing. It's not for everyone. But for homelabbers, tinkerers, and small ops teams running Proxmox in non-traditional environments, Tailmox offers something practical: flexibility. ## The verdict: worth keeping an eye on Tailmox v1.2.0 isn't some enterprise-ready migration suite. It's a clever, evolving tool that plays to the strengths of Tailscale and Proxmox—while dodging many of their rough edges. By moving to tailscale serve, it sidesteps tricky certificate binding, cleans up URLs, and reduces risk for new installs. It's not flashy, but it solves real problems with elegance. And maybe that's the bigger story here. In a sea of bloated orchestration tools and overbuilt solutions, Tailmox stays lean. It's a script, a concept, and a growing community of users figuring it out together. If you've got a homelab scattered across zip codes—or just want to experiment with remote clustering without losing your sanity—Tailmox might just be worth a spin. Just… back up your configs first.