9 posts tagged with "Headscale"

When you build an app for people running many different network setups, the hardest part is not the features. It is the front door.

Some people use Larktun's managed network. They type in an account and password, and they are in. Some prefer auth keys — a string of characters is all the identity they need. And a significant portion of users run networks that are not on Larktun at all. They operate their own Tailscale network, or maintain a self-hosted Headscale instance on a VPS somewhere.

Letting everyone walk in their own way is something the Larktun iOS app has considered from day one.

Sometimes a product does not begin with a grand plan. It begins with a small, ordinary wish.

You are away from home, and suddenly you need to publish a program, restart a service, or check the quiet little computer sitting on your desk. You only have a tablet or a phone in your hand. Still, you hope you can reach that machine as naturally as opening your own door.

Larktun grew out of that wish.

This post shows practical outcomes of transforming single-instance headscale into a multi-tenant model with isolation across tailnet, ACL, routes, DNS, CLI operations, and relay strategy.

Original post: headscale多租户改造，ACL、路由、DNS等全隔离

When scaling or rebalancing a Headscale SaaS deployment, the key challenge is how to migrate users with minimal disruption. This post presents a practical pattern: dynamic routing + system-level targeted disconnect + automatic reconnect.

Original post: headscale-系列：Headscale-SaaS-环境下的用户无缝切换实战

This post explains how to evolve headscale from a monolith into a near linearly scalable cluster using cluster_id sharding, shared PostgreSQL, an admission/control service (ACS), strict query isolation, and incremental netmap rebuilds.

Original post: headscale 系列：组建 headscale 集群，把设备接入能力做成“无限接近无限”

This article documents a practical stress-test workflow for headscale at around 200 concurrent clients, including environment sizing, bootstrap scripts, churn tests, and traffic sampling.

Original post: headscale系列：headsale压力测试

This post walks through enabling MagicDNS in headscale, plus practical usage of extra-records.json so clients can reach devices by hostname instead of memorizing raw IP addresses.

Original post: headscale系列：如何在headscale中使用MagicDNS

The official Headscale image is usually around 80MB in docker images. In contrast, custom images built from a handwritten Dockerfile are often much larger.

This post documents a closer-to-official approach: use ko to package Headscale into a container image, instead of maintaining a Dockerfile.

Original post: Headscale series: package source-built headscale as Docker image

This post is adapted from my personal blog and documents a practical workflow to build Headscale from source on Windows + WSL2 + Ubuntu, including common build errors and fixes.

Original post: Using Headscale Source Code to Build and Package