Live data from Hacker News

Replacing Systemd with OpenRC in Debian

danielcordova.me

111–120 of 168 posts

Re: Replacing Systemd with OpenRC in Debian

#111

This makes the mistake of confusing systemd the service manager and systemd the project. This is an easy mistake to make given they have the same name. The systemd service manager does not have an installer, systemd-sysinstall is a separate tool that's part of the systemd project. > systemd already integrated it, why (age verification) systemd has not integrated age verification. All that's been discussed is a simple…

Which law though?

Which country/state/region? Does it even apply? We're talking about an open source project that is not being sold, with contributors from all over the world.

Re: Replacing Systemd with OpenRC in Debian

#112
post #101

Earlier quoted context omitted.

Writing a basic init script is less intensive than having to learn the entire "schema" for both script formats, which you'd probably want to know if you were writing the generator.

Yes, but one only needs to write it once, and then everyone could use it. It could probably even be packaged as an official Debian package.

It's wild to see people having this argument in the subjunctive in 2026. I've had a convert-systemd-units tool since 2014, and provided Debian packages for the toolset that it is part of for around that long.

* https://jdebp.uk/Softwares/nosh/guide/commands/convert-syste...

* https://jdebp.uk/Softwares/nosh/guide/converting-systemd-uni...

* https://jdebp.uk/Softwares/nosh/debian-binary-packages.html

Re: Replacing Systemd with OpenRC in Debian

#113
post #54
post #43

I have a curious question. My local setup has worked for me for ages ever since arch decided to switch to systemd. Same on the servers I deal with, after Debian's switch. At the same time, I can say I'm not involved with inner workings of a Linux system enough, to be affected by init system change and the pain it might bring. In other means consider me an average Joe of the Linux world. Hence this question: If it suc…

It doesn't suck, people are just emotional beings and have some "football team" level takes on technical stuff as well. The very same people that hate systemd for "being a monolith" and limiting choice are usually also love X and hate Wayland where they can manage to explain how being a monolith is suddenly good. Especially that systemd is pretty modular - at least the actual systemd program running as PID 1. It also…

The biggest advantage systems has is also it's biggest disadvantage: the amount of functionality. Even with lot of functionality being optional this days, dlopened at runtime if present, it's still bigger then a simple shell script or.tpur own 20 lines of C-code. The later may, in some cases like running anyways only one app in full screen, more then enough. In some embedded setups you may not even need that but just run your app as pid 1 and be fine.

Re: Replacing Systemd with OpenRC in Debian

#114
post #70
post #66

Earlier quoted context omitted.

Its an argument about whether you want an init that is just an init, or an init that is designed as an extra layer of the OS. If they were comparable (i.e. if systemd was just an init system) there would be no debate. it does not help that the systemd devs are obnoxious and dismissive of concerns about compatibility. Lots of "its your problem" in their bug tracker on github.

Thanks to your summary, I finally understand. So it's about whether it's a small tool that only handles init, or whether it handles multiple layers. That's the issue. Thank you. I didn't know why there was a debate since the size difference between them is quite significant. so I don't know much about the Linux side. Have a good day.

A Linux system is build around many tools. Systemd replaced the init + gradually many othe parts, it can now handle logs, schedule jobs, manage containers, be the local dns resolver, etc. All the bricks that makes a linux system without systemd still exist. All those additional functions of systemd are optionnal but afaik they depend on systemd to handle the init to work. This is the reason openrc is often seen as a replacement as you first need to change the init to replace other functionnalities anyway.

Re: Replacing Systemd with OpenRC in Debian

#115
post #111

This makes the mistake of confusing systemd the service manager and systemd the project. This is an easy mistake to make given they have the same name. The systemd service manager does not have an installer, systemd-sysinstall is a separate tool that's part of the systemd project. > systemd already integrated it, why (age verification) systemd has not integrated age verification. All that's been discussed is a simple…

Which law though? Which country/state/region? Does it even apply? We're talking about an open source project that is not being sold, with contributors from all over the world.

IIRC laws in some US states will require OS's to collect the age of the user so services can age restrict, it does not mandate any kind of ID verification.

It doesn't matter it's an open source project, ultimately it ends up in commercial products which do require following the rules.

Re: Replacing Systemd with OpenRC in Debian

#116
post #106

This makes the mistake of confusing systemd the service manager and systemd the project. This is an easy mistake to make given they have the same name. The systemd service manager does not have an installer, systemd-sysinstall is a separate tool that's part of the systemd project. > systemd already integrated it, why (age verification) systemd has not integrated age verification. All that's been discussed is a simple…

The expressed intention behind adding the DoB field is to comply with age verification laws. Intent matters a whole lot here.

> comply with age verification laws

Do you want the developers to ignore the law and in the case of companies like Red Hat etc be vulnerable to law suits?

Re: Replacing Systemd with OpenRC in Debian

#117
post #52
post #40

Earlier quoted context omitted.

And what are your thoughts on deploying software to a fragmented system?

Systemd does not solve the deployment problem, and will not unless it adds something like a systemd package manager. It is interesting that Linux is far more widely used than alternatives that are not fragmented (e.g. FreeBSD) and has not standardised on one distro. Different people have different needs and preferences. People using Debian, Alpine, and NixOS are unlikely to agree on what they want.

> Systemd does not solve the deployment problem, and will not unless it adds something like a systemd package manager.

It really doesn't need to provide a package manager. Systemd already basically has all the knobs for launching services in their own little world (`RootImage=`, `RootDirectory=` and recently `RootMStack=` (for a de-OCI'd layered container images)) or systemd-portabled. There are probably more things I forgot.

Re: Replacing Systemd with OpenRC in Debian

#118
post #106

Earlier quoted context omitted.

The expressed intention behind adding the DoB field is to comply with age verification laws. Intent matters a whole lot here.

> comply with age verification laws Do you want the developers to ignore the law and in the case of companies like Red Hat etc be vulnerable to law suits?

There's multiple laws, and they all conflict with each other.

They are obeying some by violating some others. In this case I'd personally do nothing.

Re: Replacing Systemd with OpenRC in Debian

#119
post #111

Earlier quoted context omitted.

Which law though? Which country/state/region? Does it even apply? We're talking about an open source project that is not being sold, with contributors from all over the world.

IIRC laws in some US states will require OS's to collect the age of the user so services can age restrict, it does not mandate any kind of ID verification. It doesn't matter it's an open source project, ultimately it ends up in commercial products which do require following the rules.

The commercial products can patch it then if they so wish.

Re: Replacing Systemd with OpenRC in Debian

#120
post #109
post #43

I have a curious question. My local setup has worked for me for ages ever since arch decided to switch to systemd. Same on the servers I deal with, after Debian's switch. At the same time, I can say I'm not involved with inner workings of a Linux system enough, to be affected by init system change and the pain it might bring. In other means consider me an average Joe of the Linux world. Hence this question: If it suc…

It traded one set of problems for another set of problems. The people who are happy now don't complain (or are too new to init systems) and the people you hear are the ones who now have more problems than before. I suppose the majority either shouldn't comment because they never used anything before it to some degree, or they're just more, or they're really happier, good for them. I count myself as part of the latter…

I ran into issues early on, I used (and needed) an 'expect' script to run as part of a daemon startup, and that just didn't work with systemd. And then on another computer 'udev' spiked to 100% CPU after Debian switched to systemd, and I never managed to find out what on earth triggered that. I didn't want to spend a lot of time to solve these things, so I simply switched all my computers from Debian to Devuan (aka Debian without systemd) and I've had zero problems with that switch, which happened years ago.
Post reply on HN