The transition from Debian to systemd has been one of the most dramatic episodes in the history of GNU/Linux. The defenders and detractors of the mentioned init within the Debian community got involved in a more than fierce debate loaded with deep technical details, but that from the surface it could be understood as “standards” as opposed to the “Unix philosophy”.
After a bitter debate that led to the resignation of some outstanding members of Debian, as much in favor as against systemd, the init mainly impelled by Red Hat emerged as the winner. This decision had two consequences: the claudication of Canonical and the appearance of Devuan, a distribution impelled by those members of Debian that do not agree with systemd.
Five years ago we wondered if systemd had stayed in Debian in a definitive way
since the wounds were falsely closed and there were still many members of the community that continued not agreeing with the use of this init. And yes, so falsely were the wounds closed that recently there has been a new vote to determine if in Debian it would be better to support diverse inits or to be closed in systemd, having been the following options:
- Option 1, proposal F: Focus on systemd
- Option 2, proposal B: systemd but we support the exploration of alternatives
- Option 3, proposal A: Multiple startup support is Important
- Option 4, proposal D: Support for systems other than systemd, without blocking the feed
- Option 5, proposal H: Support for portability, without blocking the feed
- Option 6, proposal E: Supporting multiple startup systems is a Requirement
- Option 7, proposal G: Support for portability and multiple implementations
- Option 8: Deepen the discussion
The winning proposal in the vote was B, which was presented by the current leader of the project, Sam Hartman, and seeing that this is what will be applied from the future version of the distribution, we pick up on what he says:
Making use of the ability given to it by section 4.1 (5) of the Constitution
The project issues the following statement describing its current positioning with respect to startup systems, multiple startup systems, and use of systemd utilities. This statement describes the positioning of the project at the time of its adoption.
The positioning of the project may evolve over time without the need for new general resolutions (GR). The GR procedure will continue to be available if the project needs a decision and cannot reach consensus.
The Debian Project recognizes that systemd service units are the preferred configuration for describing how to start a daemon or service. However, Debian continues to be an environment where developers and users can explore and develop different starting systems and alternatives to the systemd features.
Those who are interested in exploring such alternatives must provide the development and packaging resources necessary to do that work. Technologies such as elogind, which make it easy to explore alternatives while running software that depends on some systemd interfaces, remain important to Debian.
It is important that the project supports the efforts of developers working on such technologies when there is a match between them and the rest of the project, for example by reviewing patches and participating in discussions on a regular basis.
Packages should include service units or startup scripts for starting daemons and services
Packages may use any systemd utility at the discretion of the package maintainer, as long as they do so in a manner consistent with other requirements of the policy manual and with the traditional expectation that packages should not depend on experimental or unsupported (in Debian) functionality from other packages.
Packages may include support for alternative starting systems in addition to systemd and may include alternatives for any specific systemd interface they use. The maintainers use their normal procedures to decide which patches to include.
Debian is committed to collaborating with derivative distributions that make different choices of bootable systems.
As with all our interactions with derivative projects, the maintainers involved will work with those projects to understand which changes make sense to incorporate into Debian and which changes remain exclusively in the derivative distribution.
The position defended by Sam Hartman is relatively intermediate, since in spite of not closing the door to the support of alternative inits, systemd will remain the reference. Sometimes the intermediate decisions are not to anyone’s liking, and all sorts of opinions are beginning to be seen on the net, with some interpreting this as a blow to systemd and others saying that deep down nothing will change and that this could lead to additional work to properly support alternative inits.
In short, Debian is staying with systemd, but without closing the door.