OpenSUSE: Not Everyone Likes SystemD
Phoronix: OpenSUSE: Not Everyone Likes SystemD
Not everyone is fond of migrating from SysVinit to systemd as the prominent Linux init daemon, with there being many vocal openSUSE users wanting to stay clear of systemd...
It seems to me that everyone is going to systemd because they don't wanna to maintain their own init stuff. Makes sense.
However, this push is going through while systemd is far from perfect, arguably worse than system v init style.
It's bad enough that people are willing to support their own forks.
In a more or less funny way, it's actually similar to pulseaudio. Me too, I liked the pulseaudio features and decided to close my eyes on the inherent design issues. Few years later, I'm using pure alsa, because pulseaudio design issues and complexity just annoy the hell out of me.
Systemd has design issues. The log handling. The abstraction of all the things. Binaryzation of all the things. This is far from KISS.
Are these through the eyes of a developer, maintainer, or "mom or dad using the computer" type of user?
Originally Posted by balouba
I wouldn't say classic init scripts are kiss either... they just look like it. But actually, parsing these startscripts is a lot of (unnecessary) work done hidden from the user, while only offering little.
I don't want to sound like ass but just shut the eff up and use Slackware. Or maybe port launchd instead of duplicating effort. I said my peace
Care to elborate? The binary log thing is indeed a bad idea but what else has systemd binaryzied? And where does it use too much abstraction?
Originally Posted by balouba
I do understand why people may dislike the idea of systemd inclusion, Lennart Poettering has shown time and again that what he thinks is right is the only way to go and you can go f*ck yourself if you want to use your computer differently (for instance doing things the old SysVinit way).
The Open Source movement is a major f*ck up - there are a lot less people working on Linux yet people cannot even come to an agreement and work for the same purpose of making a better system.
Those two feature requests you linked two would probably affect less than 0.1% of the Linux population. I suppose you have two choices:
Originally Posted by birdie
1. Stick with SysVinit.
2. Add those features yourself to SystemD
You can't really fault Lennart Poettering for not adding a feature that would only benefit a very very small portion of the users.
1. How, if most distros will switch to it sooner or later as user space components become dependent on it??
Originally Posted by gururise
2. Are you crazy? Lennart Poettering can as well say, "go f*ck yourself and maintain your patches separately forever" (Ulrich Drepper also from RedHat is notoriously known for that).
I love Open Source. I really do.
And BTW, systemd itself can leave your system unbootable (!!) - go ask Fedora 18 testers - they will surely demonstrate to you their anger.
A technical reason why systemd sucks big time?
Try to boot your system the old init = /bin/bash way, you won't be able to do anything with it, as udev was replaced with systemd and systemd cannot run when it's not a system init process.
That's called broken by design.
Sorry, but I have to say, f*ck Lennart Poettering and f*ck systemd. systemd has become a single point of failure, a Swiss army knife which once getting broken renders your system completely unusable. That's f*cking wrong. That reeks of bad incompetent design decisions.