"Boycott Systemd" movement takes shape

by
in linux on (#2S4F)
story imageSome people have had enough, and they've organized a boycott at "http://boycottsystemd.org" to organize efforts. From the top: "Disclaimer: We are not sysvinit purists by any means. We do recognize the need for a new init system in the 21st century, but systemd is not it." OK, that's enough to keep me reading. They outline twelve well-thought-out reasons systemd is dangerous, and a set of ways you can get involved, including refusing to use systemd distros, moving to slackware, crux, gentoo, BSD, and more. Here's just one of them:
systemd clusters itself into PID 1. Due to it controlling lots of different components, this means that there are tons of scenarios in which it can crash and bring down the whole system. But in addition, this means that plenty of non-kernel system upgrades will now require a reboot. Enjoy your new Windows 9 Linux system! In fairness, systemd does provide a mechanism to reserialize and reexecute systemctl in real time. If this fails, of course, the system goes down. There are several ways that this can occur9. This happens to be another example of SPOF.
Interesting times. When's the last time you heard someone advocate moving immediately to Slackware or Gentoo?

Re: F Word (Score: 1)

by caseih@pipedot.org on 2014-09-11 02:54 (#2S9C)

Except that init scripts have always been very fragile and complicated. In short they've never worked that well. They are full of hacked logic to check for running instances using all manner of devices (PID files, grep for the process name in ps, check /proc), they don't lend themselves to parallel execution, and they have no facilities whatsoever for process supervision. Sure you can hack a cronjob (and many a sysadmin has done that!), or use a third-party supervisor daemon (who watches the watchers?). Every commercial unix out there has ditched system V init scripts for these reasons. For example, Solaris hasn't had init scripts ever since maybe Solaris 9? Apple ditched init scripts in OS X quite a long time ago. 10.4 I think?

Systemd isn't that complicated. It's highly modular also. It does a good job and it is remarkably stable for such a new project. Rather than hacking a complex init script, I can create a service in just 3 or 4 lines of a simple, clear, conf file now. In fact it dramatically simplifies the development of custom services. I don't even have to do the fancy fork dance to put a daemon in the background. From a sysadmin point of view, this sort of thing has been sorely lacking in Linux. If I feel inclined, I can integrate with systemd's socket management system, and get instant, transparent parallel startup capability. But the same daemon can still run on non-systemd machines (if I actually implemented daemonization myself). I'm hard pressed to find any negatives, really. And this is coming from a guy who had nothing good to say about systemd before I actually used it. Seems like most of the FUD comes from people like me, folks who've not interacted with it much at all. It's not the only game in town, but it certainly works, and I have no problems with major distros standardizing on it.

On a related note, I would not want to live without pulseaudio either. I currently use pulseaudio regularly to do things that were never possible with ALSA before, such as directing certain program's outputs to certain audio devices. Very handy indeed.

You can boycott systemd if you want, of course. Anyone is free to roll their own distro if they feel like their distro of choice has been bought out by some nefarious power. There are several alternative init systems to choose from.
Post Comment
Subject
Comment
Captcha
Tomorrow is Tuesday. If this is true, what day is today?