>>6 First, who are you quoting? Second, systemd is bad both because of Poettering, and because it also has awful design principles.
Name:
Anonymous2017-12-29 7:38
First off, systemd is not an init system, it has an init system as part of the systemd suite. systemd is a project to build a standardised lowlevel userland for Linux. The project is pretty comprehensive and it delivers a lot of functionality under one umbrella. It does away with a lot of older, often undermaintained software packages, which were traditionally used to assemble a low level userland.
Which is where the contention comes from, as a system suite systemd is restrictive for Unix virtuosi who are used to tailor a system with wit, ingenuity, a lick and a prayer and a couple dozen of unrelated packages. systemd makes such knowledge useless.
The faction that thinks that systemd is Linux's Hiroshima, finds all the added functionality bloat, unnecessary and dangerous, as it is all under development in one project.
All the systemd jokes stem from the comprehensiveness as a low level system suite. People against it love to joke that one day systemd will write its own kernel.
There is a lot of FUD and hate going around. Some arguments do have merit, a lot of eggs in one basket is certainly true, but as with all things in life, it depends which tradeoff you prefer. Do you want a suite of well designed software, working closely together, so that system management is streamlined or do you want the complete freedom to tailor your own low level system with a lot of time tested, interchangeable components.
I have no desire to be a low level system designer, so I prefer systemd. I don't hate traditional init systems though. If a Linux system has one and I need to work with it, I'm still happy it boots and starts the necessary services.
Name:
Anonymous2017-12-29 20:58
Lmao how the fuck you make (post truncated.) long poast based on my shit thread? >>8
30% of Russian military is Muslim. Russia has open borders for nigs and asians. Russia has some fag called Dugin sperging out about huwhites and capitalism like a typical kike. Must I continue to go on?
>>8 OpenRC has the supposed ``improvements'' that systemd brings. The major Linux distributions could have chosen that instead of wasting time trying to turn the foul diarrhea garbage that is systemd into barely tolerable acidic urine.
>>14 Apples an oranges. I think people are mad moreso because they can't think of systemd in terms other than Another Init System. If there were a counterpart to systemd whose scope is just as large, I wonder how people would react? Because, despite what the more polite, mature side of the debate says, I don't think it's entirely pettiness. A good reflection of that is pulseaudio. And I wonder how people will see pulseaudio as a competing solution, pipewire, becomes mature enough to be actually used.
I think people are mad moreso because they can't think of systemd in terms other than Another Init System.
Nor should it be. It's over engineered, and does much more than is needed, and instead of helping, it just shits all over the place. It's only in place simply because of politics.
Name:
Anonymous2017-12-30 19:13
CHECK EM
Name:
Anonymous2017-12-31 14:48
>>16 Systemd is popular because it solves a real problem of init configuration. Openrc wasn't good enough when there were decisions made to change to Systemd.
Name:
Anonymous2017-12-31 15:02
>>16 If by politics you mean practicality, then yes. It might not be apparent as a desktop user, but systemd makes it significantly easier to deploy multiple specialized distributions in enterprise by means of uniformity. And before you say that's an oxymoron, there's much more to a distribution than just systemd. If you don't believe me, just compare Ubuntu to Exherbo.
Systemd is popular because it solves a real problem of init configuration.
Which was exactly? The only ``benefit'' I see from it is parallel service startup, but OpenRC already has that. >>19
systemd makes it significantly easier to deploy multiple specialized distributions in enterprise by means of uniformity.
You could already do that long before systemd was around.
there's much more to a distribution than just systemd.
No kidding. And, yes, implementation of systemd was done by politics, not because of any ``improvement'' over SysVInit. And don't get me wrong, I'm not opposed to modernizing aspects of the init system, but systemd just takes Linux development backwards and people wasting time trying to improve on that instead of taking on something that has much less complexity, bloat, and other nonsense.
You could already do that long before systemd was around.
Except you couldn't. You try to deploy Gentoo on the servers of a predominantly Suse environment, and auditors would have your ass, not just because of their unfamiliarity with Gentoo and the superstitions that come therewith, but because they would have a genuine, logical argument as to why you can't do such a thing: uniformity. Different package names, different scripts, different release cadences.
You try to deploy Gentoo on the servers of a predominantly Suse environment, and auditors would have your ass, not just because of their unfamiliarity with Gentoo and the superstitions that come therewith, but because they would have a genuine, logical argument as to why you can't do such a thing: uniformity.
Why would you switch to a different distribution if you have competent sysadmins to fix things when they break?
Name:
Anonymous2018-01-01 5:30
>>25 Do I really have to spell it out to you why you might deploy Suse on desktop and Gentoo on server?
>>26 What's stopping you from using one distribution for both? Especially if your aim is ``uniformity'' across everything?
Name:
Anonymous2018-01-01 11:28
>>28 >>25 My point was that there is more to a distribution than just the init system and whatever the hell else systemd encompasses. There are startups out there with ultra-minimal, optimized Exherbo servers. And while Exherbo may support runit + sinit/sysvinit/s6-linux-init + openrc/s6, it supports systemd the most simply because it's the most well-supported, mutually-liked, economical of all the options. Conversely, you could never deploy Gentoo on desktop because, even with binhost, it would be blatantly inefficient, nor would the Gentoo Project support you in the same way that Redhat would. The distinctions between the init systems and the lumps and discrepancies that exist for scripting with each one of them is of no practical value--it means more work, which isn't a reflection of skill, as you seem to be implying but just another vector for failure.