100,000 Websites on One Linux Server
Hotserv.dk was a free web hosting platform I built when hosting was still expensive. Custom-compiled Apache, PHP, and MySQL on a single server running 100,000+ sites.
In 2005, web hosting was expensive. A basic shared hosting plan cost real money for a teenager who wanted to learn PHP. I built Hotserv.dk as a free hosting platform where users got a subdomain or could point their own domain, with no ads. The whole thing ran on a single Linux server.
The numbers
100,000+ websites on one machine. This was only possible because I custom-compiled Apache, PHP, and MySQL specifically for this workload. Every configuration setting was tuned for the pattern of thousands of small sites sharing resources. Memory limits per process, connection pooling, filesystem optimizations, process recycling. Default configurations would have collapsed under the load.
The community
A community formed around Hotserv. Users learning PHP for the first time, building their first websites, asking for help in the forums. I recruited volunteer moderators to help manage the growing user base. Hotserv became a place where a generation of Danish developers wrote their first lines of code.
That was the part I didn't expect. I built a hosting platform. What I got was an educational community.
Why it ended
Two reasons. First, hosting prices dropped dramatically. When you can get shared hosting for a few kroner a month, free hosting loses its appeal. Second, dealing with abuse became a constant, draining fight. Users uploading illegal content, phishing pages, malware distribution. Policing 100,000 sites is a full-time job I didn't sign up for.
The same pattern hit Frac.dk, a similar free hosting platform started by Soren Nielsen in 2003 that I had joined earlier. Free platforms attract abuse. It's an unavoidable cost.
What stayed with me
Running 100,000 sites on one server taught me more about Linux performance tuning than any course could. Memory management, process limits, filesystem optimization, connection pooling. These days I use Kubernetes to manage containers at scale. The fundamentals are the same: understand your resources, tune for your workload, monitor everything.