Twelve Years of Yelling: A Field Guide to the Init War Ceasefire

There was a time — roughly 2014, give or take a kernel release — when you couldn't mention systemd in a Linux forum without the thread descending into something resembling a homeowners association meeting. People who agreed on 99% of their computing philosophy would reduce each other to ASCII rubble over an init system. It was glorious. It was exhausting. And sometime in the last couple of years, it mostly just... stopped.

The Great Init War, it turns out, didn't end with a winner's speech. It ended the way most long arguments do: everyone got old, everyone got busy, and systemd shipped anyway.

How We Got Here: A Brief History of Yelling

For the uninitiated (pun absolutely intended, we're not sorry): traditional Linux used init systems like SysV init — a set of shell scripts that started services in a loosely defined order. It worked the way a paper map works. It was also, its critics argued, slow, fragile, and about as parallel as a single-lane road in Ohio.

systemd arrived with a pitch that was hard to argue with: faster boots, parallel service startup, dependency tracking, cgroup-based resource management, and logs you could actually query. Lennart Poettering and Kay Sievers built something genuinely modern, and modern things in the Linux world are treated with roughly the same suspicion as a stranger offering you a ride.

Then came the flame wars. The Debian forks. The mailing list threads longer than some novels. The famous incident involving the Fedora community and enough hot air to warm a small city. People wrote open letters. People left projects. At least one person famously deleted their entire online presence over the argument. For a while, "systemd or not" was the shibboleth that determined which corners of the internet you could safely stand in.

Where the Holdouts Are Now

The anti-systemd resistance didn't vanish — it just moved to quieter neighborhoods.

Devuan, the Debian fork born specifically to escape systemd, has settled into a respectable middle age. It's stable, it's maintained, and its community has largely traded apocalyptic rhetoric for the quiet satisfaction of people who got out of the argument alive. Its user base isn't large, but it's loyal the way people who've survived something together are loyal.

Void Linux runs runit, an init system so minimal it makes SysV look baroque. Void's pitch isn't ideological so much as aesthetic: small tools, done well, no drama. It has attracted a following of people who describe themselves as "just tired," which may be the most honest marketing in open source.

Gentoo remains Gentoo — a place where compiling your own kernel is a personality trait, and choosing an init system is simply another weekend-long configuration project. OpenRC is alive and well there, and Gentoo users will happily tell you about it for as long as you can stand.

And then there's the broader ecosystem. Most mainstream distros — Fedora, Debian, Ubuntu, Arch, SUSE, Red Hat — shipped systemd years ago and haven't looked back. Even the corporate world, which once viewed the whole fight as a weird hobbyist scuffle, has standardized around it. If you run a container in a managed cloud service right now, statistically speaking, systemd probably started it.

The Veterans Weigh In

We talked to folks on both sides of the old divide, and the responses were strikingly similar.

A longtime sysadmin who spent years on the anti-systemd side put it this way: "I still think it does too much. But I also haven't thought about my init system in two years, which is arguably the highest compliment you can pay infrastructure. It's like a good waiter — you only notice when it's wrong."

A distro maintainer who fought for its adoption was more blunt: "The people yelling about scope creep were right about the philosophy and wrong about the outcome. Turns out users don't want 40 small tools; they want their laptop to boot and their logs to be findable. Both things can be true."

Even the most passionate holdouts concede a ceasefire. "I run runit on my personal boxes and I don't miss systemd," one Void user told us. "But I also don't miss the arguing. I'd rather spend that energy literally anywhere else."

Could That Energy Have Built Something?

Here's the uncomfortable question the ceasefire leaves behind: what if all that effort had gone into, say, documentation? Or mentoring? Or a competitor that was better instead of just different?

The counterfactual is tempting. Thousands of person-hours went into forks, rants, and forum threads that produced no code. The Devuan team, to their credit, actually built something — which is more than can be said for most participants in the discourse. But the broader lesson stands: outrage is a terrible project management methodology.

At the same time, the critics weren't wrong about everything. systemd's scope really has grown. The project really does embody a philosophy of integration that sits awkwardly with the Unix tradition it grew from. The difference is that the community has largely decided those are acceptable costs, and no amount of arguing has moved that needle.

So Is the War Over?

Mostly. There will always be a fresh thread on some forum where someone discovers journald's binary logs for the first time and experiences the stages of grief in real time. But the organized resistance has demobilized, the forks have found their niches, and the mainstream has moved on to new arguments — Wayland, Rust in the kernel, whether AI-generated code counts as contribution.

The Init War's real legacy might be its ending: proof that the Linux community can have a genuinely divisive, decade-long fight and come out the other side with the ecosystem intact, both sides still shipping, and everyone a little more tired and a little more humble.

systemd turned twelve this year. It's a teenager now — moody, sprawling, everywhere, and impossible to imagine the household without. The yelling has stopped. The daemons, as always, keep running.

Agree? Disagree? Ready to restart the whole war in our comments section? Honestly, we respect the commitment either way.