Hey guys, I have been using a Linux for a while and I always thought as long as some software is Open Source, its good enough.
What are other considerations that make for good FOSS software? In which of these does systemd fail? Is that why it’s criticized or is there some other reason?
FOSS being good enough is the baseline, not the finish line. Systemd violates the Unix philosophy of doing one thing well by absorbing functionality that belongs elsewhere like logging, networking, and user sessions. Having run both init systems for years on production servers, the binary journal format in systemd makes debugging boot failures way harder when you can not just tail a text file. Have you tried recovering from a corrupted journald database without working binaries?
There’s like a million threads on this already. If you’re genuinely interested you would have looked, but you’re clearly karma farming for a brand new account.
Come on, that’s pure ragebait… It’s not like it"s hard to search for all the different reasons… https://nosystemd.org/ collects a few in semi-coherent form. The short answer is: It combines a lot of previously independent systems responsibilities under one umbrella organization that holds decidedly strong opinions and is not exactly open to criticism.
I don’t think there is a problem with it. It’s a piece of software that people can choose to use if it fits their specific use case. It has a long list of features and abilities, and lots of people find it sufficient for what they want.
There are also a lot of people that don’t find it suitable for what they want and they can choose to not use systemd and use some other options.
I personally don’t use systemd. I have used it for a while, originally I used sys v for a long time, then arch adopted systemd, I tried to get used to it and understand it but never felt comfortable with it, so I moved to void Linux which uses runit plus other items to replace systemd, and I feel a lot more comfortable and happy with this.
You do your research and testing and find what fits your use case.
Probably rage bait.
Anyway, I like the simplicity of OpenRC and prefer it over systemd. After many years, I never had a reason to switch my Gentoo boxes (servers, laptops, etc) over.
I use systemd at work too, and it always feel unnecessarily complicated and that fixes issues I never had.
Whatever fit your bills I guess. Choice is good, so I am happily applying my choices with OpenRC.
The main problem is that people don’t understand that it is NOT “just an init system”
No, the main problem is that it’s NOT just an init system.
Mixing up different concerns is just bad design.
systemd is only used to initiate booting of a system afaik and, in the past, it was criticized for overreach (as in it did more than its predecessor did).
now-a-days, it seems to be criticized for pre-complying with big brother like laws like age verification.
Ahem, it does a ton more than merely “initiate booting” (logging, time, user management, device management, the list is long and it is really hard to find a piece of basic system functionality it hasn’t subsumed), please don’t spread misinformation.
you must have missed this part:
it was criticized for overreach (as in it did more than its predecessor did).
Nope, but “more” doesn’t adequately capture the scale of it. Given that you talk about a predecessor (singular) I presume systems programming is not your specialty?
op was asking about linux specifically and neither sysvinit, upstart, openrc, or runit did more than booting
No shit. “More” is technically correct. “A metric fuckton more” conveys the appropriate scale. It’s not like systemd added a handful small features, it has subsumed nearly everything between kernel and userland. (Note that proponents usually point this out as a good thing; It’s uncontested, you just seem blissfully unaware)
- It came from a corporation so was designed by 1 guy not a committee and some people will never forgive it for that.
- The systemd suite is far more than an init system and keeps getting bigger, I genuinely think it’s just a matter of time until it has a mail retrieval service built into it.
- It offers clear benefits so even distros resisitent to depending on a tool when there are alternatives have adopted it
- It doesn’t pander to slacktivist on stuff like including an optional DoB field
That’s about it, I find it kind of annoying sometimes as it messes with stuff I knew how to do (harden per-user-tmp partitions), but overall the benefits to distro maintenance must be worth it so I don’t worry about it too much.
Much like the DoB stuff I find the pushback to it far more annoying than the actual inconvenience of sometimes not being able to configure a tweak how I’d like.
The reason this person is getting downvoted with no answer to their good points is an indicator as to why the subject is sensible.
Someone (admittably an asshole) found how to solve a problem, wrote a suite that does a very good job where the alternatives were absolute crap, and all major distros started using it.
Everybody is free to write another alternative, but everybody that is capable knows that systemd is probably better than whatever they could make, so that is what almost everybody is using.
Have you considered that we don’t think any of those are “good points” worthy of comment?
