Arch Linux Is Switching To Systemd

[ArchLinux]

I just hope thats the real reason the switch is being made.

Why wouldn't it be? It also eases up the maintenance of other packages by absolving them from things like ck. Tomegun also brought up some points in the forum, as well as Allan in his "Are We Removing What Defines Arch Linux?" blog post.

[ceage]

5. I don't even understand what the hell you are trying to say here...

He's saying that his issues with systemd are purely emotional. That's why there's no technical argument anywhere in that quoted post. This is probably also true for 99 percent of all the other criticism voiced in connection with systemd on websites.

The real arguments are generally brought up on systemd-devel and discussed by people who have some idea what they're talking about. So far, I've never read anything on that list along the lines of systemd being a broken concept. So I'd say there isn't that much to worry about. (The non-portability argument is not even worth discussing, since open code is inherently portable.)

[ChrisXY]

I don't want an init system by a guy that wrote a real time daemon, dropped into a jack ml and told them 'btw, this will be mandatory, and yes, I screwed you over. Sucks to be you'.

In what way is it mandatory?

[gururise]

Wow, so much FUD in this thread, just like all the threads in the arch-general mailing list.

A. You can still use rc.conf, some of the Arch devs have worked hard to make their systemd be able to read the DAEMONS array so no changes necessary, though they are recommended to match upstreadm.

B. The main selling point of systemd is not speed, though it is much faster. The reason the devs are all so eager to switch is because systemd service files are MUCH easier to write than initscripts. Not only are they much faster to write but they're portable so the hope is upstream will eventually be able to maintain their own service files and the devs jobs will be much easier. People do things that make their lives easier - surprise!

C. They didn't force you to install systemd, systemd-tools is a collection of small binaries which are useful to any init system - Arch's initscripts make heavy use of them. AFAIK they're still not forcing you, but it will be default and they probably will force you eventually.

Thank You. I was wondering why switch from initscripts to systemd if the old method wasn't broken. Your post, seems to indicate the major advantage of switching to systemd is simplification of the boot process. I guess if thats true, then thats as good a reason as any to switch to systemd. I just hope thats the real reason the switch is being made.

[bwat47]

I don't want an init system that was only created because pulseaudio sucks.

I don't want an init system by a guy that wrote a real time daemon, dropped into a jack ml and told them 'btw, this will be mandatory, and yes, I screwed you over. Sucks to be you'.

I don't want an init system that depends on dbus

I don't want an init system that gnomifies my computers.

I don't want an init system that replaces a simple init 3 with this crap: systemctl enable multi-user.target

what next? mandatory initrd even on a system where no mounting is needed for all the important stuff to be in place?

---- energyman76b gentoo.org forums

Sums up my thoughts on SystemD 100%

This post is so incredibly moronic, I don't even know where to begin.

This video sums it up nicely:

http://www.youtube.com/watch?v=5hfYJsQAhl0

anyway:

1. How the fuck was systemd 'created because pulseaudio sucks'? That literally makes no sense, pulseaudio and systemd have absolutely nothing to do with eachother, besides being created by the same author.

2. Again, what the fuck do rkit and pulseaudio have to do with systemd? NOTHING.

3. Why is this bad exactly? Dbus is extremely useful and tons of software already depends on it.

4. How the flying FUCK does it "gnomify" your computer? That makes absolutely no sense, systemd has nothing to do with gnome.

5. I don't even understand what the hell you are trying to say here...

[finalzone]

I don't want an init system that was only created because pulseaudio sucks.

I don't want an init system by a guy that wrote a real time daemon, dropped into a jack ml and told them 'btw, this will be mandatory, and yes, I screwed you over. Sucks to be you'.

I don't want an init system that depends on dbus

I don't want an init system that gnomifies my computers.

I don't want an init system that replaces a simple init 3 with this crap: systemctl enable multi-user.target

what next? mandatory initrd even on a system where no mounting is needed for all the important stuff to be in place?

---- energyman76b gentoo.org forums

Sums up my thoughts on SystemD 100%

Good. Welcome to the extinction terrority where those who refuse to adapt die.

[Rallos Zek]

I don't want an init system that was only created because pulseaudio sucks.

I don't want an init system by a guy that wrote a real time daemon, dropped into a jack ml and told them 'btw, this will be mandatory, and yes, I screwed you over. Sucks to be you'.

I don't want an init system that depends on dbus

I don't want an init system that gnomifies my computers.

I don't want an init system that replaces a simple init 3 with this crap: systemctl enable multi-user.target

what next? mandatory initrd even on a system where no mounting is needed for all the important stuff to be in place?

---- energyman76b gentoo.org forums

Sums up my thoughts on SystemD 100%

[bwat47]

Systemd has thus far not been a huge problem, yet I truly believe both upstart and launchd are better systems. BSD Init scripts are too limited, and they were my no. 1 reason to avoid Arch. A modern Init system is important, and anything recently written in C was an improvement. Systemd limitations aside... Although the lack of portability rubs me the wrong way.


PulseAudio is terrible, 100 ways to Sunday, superfluous doesn't even describe it's worthlessness. Even the PulseAudio developers admit, unwillingness/inability to improve ALSA was used to excuse it's dreaded conception. This actually describes, extremely accurately, what is most irritating about the FOSS development community. When something needs to be improved, instead of cleanly implementing improvements, an extra layer of latency is always the response.

Qualms about pulseaudio's implementation aside, saying its completely terrible or superfluous is just stupid. For my usage plain alsa straight up does not cut it, certain things either do not work at all or take significant configuration (for example my hdmi audio output on my laptop). Pulseaudio handles all these things perfectly for me out of the box.

[techzilla]

What must be stated

Systemd has thus far not been a huge problem, yet I truly believe both upstart and launchd are better systems. BSD Init scripts are too limited, and they were my no. 1 reason to avoid Arch. A modern Init system is important, and anything recently written in C was an improvement. Systemd limitations aside... Although the lack of portability rubs me the wrong way.


PulseAudio is terrible, 100 ways to Sunday, superfluous doesn't even describe it's worthlessness. Even the PulseAudio developers admit, unwillingness/inability to improve ALSA was used to excuse it's dreaded conception. This actually describes, extremely accurately, what is most irritating about the FOSS development community. When something needs to be improved, instead of cleanly implementing improvements, an extra layer of latency is always the response.

[fackamato]

systemd shaved off 70% off my boot time, very welcome!