Linked by dungsaga on Tue 8th Sep 2015 23:31 UTC
General Unix
The subject of process management, supervision and init(8) for Unix-like systems is one plagued by a large degree of ahistoricity and "pop culture" explanations. This leads to a lot of confusion and misunderstanding surrounding feature sets and how one formulates the problems surrounding reliable process management on Unix in general, making it a ripe topic for demagogues of all persuasions.


The purpose of this article is to set the record straight on the history of attempts to create "modern" init systems, where we define "modern" somewhat broadly as anything that tries to improve the classical BSD and System V styles of initialization and service management.

Order by: Score:
v ?
by kragil on Wed 9th Sep 2015 04:43 UTC
RE: ?
by cb88 on Wed 9th Sep 2015 04:53 UTC in reply to "?"
cb88 Member since:

You don't have to hate change or have a neckbeard to hate systemd.

Also, my lair is on the 2nd floor of a townhouse you insensitive clod >:D !

Edited 2015-09-09 04:54 UTC

Reply Score: 4

RE: ?
by on Wed 9th Sep 2015 12:52 UTC in reply to "?" Member since:

Did you even read the article!? Even Thom's summary has an excerpt where he states it's just a quick run through history so that we will know how we got to where we are.

I actually think it was written by someone defending systemd. When you're done reading you come out with the feeling "yeah, only systemd puts every issue the other inits systems were trying to solve into 1 system."

After having read the whole thing, I believe the takeaway from the article is: "People have been trying to find a suitable replacement for SysV init for over a decade. One finally becomes popular and people come out of the woodwork to hate it."

Personally, I like some of systemd's stated goals and dislike others. I think too many people are either 100% for or 100% against it, and we end up with the current stalemate/sh*t-throwing that we have.

Edited 2015-09-09 13:05 UTC

Reply Score: 5

RE[2]: ?
by matthekc on Fri 11th Sep 2015 04:53 UTC in reply to "RE: ?"
matthekc Member since:

In my semi-educated worthless opinion as long as the underlying ideas prove to be at minimum not a complete blunder then the little paper cut issues will get sorted.

I think systemd will be okay even with or without the current developers. There are too many people and distributions involved now pushing it in the directions that cover most usage cases. There of course will be people that never move over and that's okay too.

Reply Score: 2

by uridium on Wed 9th Sep 2015 06:52 UTC
Member since:

systemd gives me gas.

Reply Score: 2

RE: History
by avgalen on Wed 9th Sep 2015 09:19 UTC in reply to "History"
avgalen Member since:

FREE gas!

Reply Score: 3

RE: History
by Sauron on Wed 9th Sep 2015 12:51 UTC in reply to "History"
Sauron Member since:

systemd gives me gas.

It gives me nightmares!

Reply Score: 2

by Carewolf on Wed 9th Sep 2015 10:28 UTC
Member since:

It seems to skip what we had before SysV. It was not always there, it was just imported because it was used on the big UNiX'es so Linux had to have it too. The main benefit was having more than two options at boot, instead of single-user and normal, it had single-user, normal with X11, normal without X11. That was the whole benefit (and not used today).

Before then you would either boot to X11 and hope it worked, or you would play it safe and boot to console and then launch X with startx.

Edited 2015-09-09 10:29 UTC

Reply Score: 3

RE: Pre-SysV?
by IgnitusBoyone on Wed 9th Sep 2015 14:15 UTC in reply to "Pre-SysV?"
IgnitusBoyone Member since:

I still boot to framebuffer consoles and then type startx. To me init systems are more about shutting down then starting up. Personally, I'm a fan of the interaction model of OpenRC, but I really hate the scripting model I assume this roughly translates to "I've used OpenRC for a decade and it is the best".

Reply Score: 1

runit (2004)
by jello on Wed 9th Sep 2015 16:44 UTC
Member since:

Really sad to not see runit mentioned.
Using it on arch and like it very much.
Easy to use and set up.

'nuff said ;)

Edited 2015-09-09 16:45 UTC

Reply Score: 4

RE: runit (2004)
by bnolsen on Thu 10th Sep 2015 03:57 UTC in reply to "runit (2004)"
bnolsen Member since:

take a gander at void linux which uses runit by default. i really dislike having to fundamentally rig arch.

Reply Score: 2

GoboLinux's BootScript System
by tupp on Wed 9th Sep 2015 18:31 UTC
Member since:

The article didn't mention the efficient BootScript system used by GoboLinux since about 2002.

Reply Score: 4

by Gullible Jones on Sun 13th Sep 2015 19:25 UTC
Gullible Jones
Member since:

I'll admit to not being the biggest fan of systemd, mostly in terms of (IMO) overcomplicated nature, and also how the tools behave. OTOH, look what it brings with it:

a) Consolidation of a sizable chunk of the Linux userspace in one place... With a nice wiki, online docs, publicly accessible version control, and decent uptime. This is a much better state of affairs than having sources for vitally essential system components scattered across dozens of personal websites.

(If you think I'm kidding about the latter, check out the Linux From Scratch cookbook.)

b) A unified, standardized, and at least half-way robust init system used by almost every major Linux distribution. Freakin' hurray!


Although really, it's the harassment campaign against Lennart Poettering (!!!) that has my mind made up by now. If a person can make people that uncomfortable, and attract that level of rage, just by publishing useful software, then I tend to suspect they're doing something right.

Reply Score: 2