this post was submitted on 06 Jun 2025
52 points (81.7% liked)
Linux
7708 readers
241 users here now
A community for everything relating to the GNU/Linux operating system
Also check out:
Original icon base courtesy of lewing@isc.tamu.edu and The GIMP
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
No...systemd was controversial because it complicated an entire ecosystem and caused lots of growing pains for very little payoff at the time. SysV was fine for many, but now so is systemd, and it's solved many growing pains for distro maintainers.
@just_another_person @rumschlumpel The idea of replacing system-V init with an init system capable of parallel start-ups in an era where multi-core CPUs became the norm makes sense. If it had stopped at this I would have been fine with it.
But it then goes and takes over DNS and in a way that breaks some mail sites that have spf records in a single record longer than 512 bytes which is officially against the DNS standard but which bind9 was fine with, then it had to take over system time keeping, and then user home directories, and then it wants to containerize everything.
The original Unix and by extension Linux philosophy was make one tool to do one thing and make it do it well.
Systemd by contrast is now one bloatware that wants to do everything and doesn't do everything well. It does perform it's function as a new init well.
I think systemd has moved desktop and server Linux towards being more BSD-like ... and I don't necessarily think that's a bad thing.
Maybe we'll end up needing an X11 -> Wayland sort of transition where there are protocols instead of "an implementation."
However, I've yet to see systemd be meaningfully detrimental. Are distros a little less different? Yeah. Has it made my life easier when I need to go between distros? Also, yeah.
I think on some level, we're just getting to a more mature Linux desktop and server ... and as a result consolidating on stuff that really doesn't have strong reasoning to be different.
I mean, systemd-networkd and systemd-timesyncd are both completely independent and are not required by systemd. I use connman and chronyd on my arch box and systemd gives not one fuck.
There's still some totally valid concern to be had over how bundled a lot of this stuff is, but it's not all one big blob.
@Badabinski @just_another_person @rumschlumpel @propitiouspanda Yes but they are becoming the defaults on many distros. In particular systemd-resolvd is an issue because it enforces the 512 byte limit on txt records. The problem with doing this is many large sites have spf records longer than 512 bytes and fail to break them up into separate txt records, so if you enforce this limit and they initiate mail from one of the truncated hosts, it gets rejected. This is not good and so I've worked around this by disabling networkd-resolvd and installed bind9 instead. I've actually had no problem with timesync but why re-invent all the wheels? To me it seems Poettering is a control freak and wants to take over my systems.
To be honest, stuff not working when it breaks the standard is unfortunate, but I wouldn't blame this on the tool that adheres to said standard.
You're not inconvenienced by systemd-resolvd, you're inconvenienced by those mail sites doing stuff that doesn't work, possibly as a result of them needing to do something that was slightly flawed to begin with: using DNS records to possibly hold more data than they can per the spec, which, if I understand things correctly, is because of the limitations of UDP traffic.
Not that that helps you, of course, it's annoying and I recognise that.
@dnzm Some of us live in the real world where we have customers that expect to receive their e-mail and aren't interested in the details of a standard, and since prior to systemd this was not an issue, I see no benefit to making an issue. UDP packets can be any arbitrary length up to 65535 bytes (including the header), there is no sound reason for limiting them to 512 bytes.
I, too, live in this fabled real world, and I already mentioned I understand your issue. I just think you're barking up the wrong tree, but luckily you're able to work around things, and that's the most important bit, isn't it?
@dnzm Yes but why if you're going to do something to "improve" linux, and honestly the fast parallel start up IS an improvement, but why then go on to try to take over the entire operating system and enforce limits that were not enforced before, or at least why not make it an option?