Debian Tech Committee Falling Further Into Disarray

[curaga]

Yes. My objections to systemd are twofold, and as long as they stand, its excellence in other areas (such as better unit files or daemon management) cannot override. This has been discussed here many times, but heh, recap time :P

1) It does far too much in pid 1. This greatly increases the risk to system stability. "All eggs in one basket" is simply stupid.

2) The attitude and politics of its creators. Subsuming everything like a tentacle monster, actively working against loose coupling.

[TheBlackCat]

Just responding to this part: it does. Sysvinit is indeed a better option than systemd. Just because you disagree does not mean they don't know what they're doing.

That may be your opinion, but Ian had consistently ranked systemD higher than sysV until this vote.

[oleid]

recap time :P

Yeay!

1) It does far too much in pid 1. This greatly increases the risk to system stability. "All eggs in one basket" is simply stupid.

It seems to make sense to put certain stuff into pid 1, e.g. cgroup management. But as long as it's well-tested...

2) The attitude and politics of its creators. Subsuming everything like a tentacle monster, actively working against loose coupling.

Well... They always argued it makes technically sense to combine that stuff. I don't think they have an agenda to destroy variety. They, however, don't seem to really care if other solutions are incompatible to their solution, if it provides an technical advantage.

[TheBlackCat]

1) It does far too much in pid 1. This greatly increases the risk to system stability. "All eggs in one basket" is simply stupid.

What, specifically, does it do in pid 1 that it shouldn't

2) The attitude and politics of its creators. Subsuming everything like a tentacle monster, actively working against loose coupling.

Right, because systemD maintainers are kidnapping the families of developers of other projects and forcing them to merge their projects systemD. Somehow it doesn't seem like a strength of systemD, rather than a flaw, that other projects see value in joining with them.

[oleid]

[...] Somehow it doesn't seem like a strength of systemD, rather than a flaw, that other projects see value in joining with them.

Well, what choice do they have, considering the systemd developers detain their families?

[curaga]

What, specifically, does it do in pid 1 that it shouldn't

Everything besides reaping zombies. Each and every action, each and every dependency (dbus!) increases the chance of bugs. Each line of code adds complexity.

[danwood76]

Everything besides reaping zombies. Each and every action, each and every dependency (dbus!) increases the chance of bugs. Each line of code adds complexity.

I assume you don't run the Linux kernel then because there are more than a few lines of code contained in that package which could obviously cause bugs.

Personally if that is the biggest argument against SystemD you have then you should just leave the party now.

Upstart sucks and my experience of SystemD on Arch so far is that it is excellent (my opinions obviously).

[oleid]

Everything besides reaping zombies. Each and every action, each and every dependency (dbus!) increases the chance of bugs. Each line of code adds complexity.

IPC makes it possible to split the whole thing into seperate processes, so DBUS is necessary as is the cgroup stuff, which PID1 uses to keep track of child processes. You absolutely can't exclude DBUS or the cgroup stuff. And besides that, AFAIK there is not much in PID1.

[DDF420]

What, specifically, does it do in pid 1 that it shouldn't

systemd Broken by design http://ewontfix.com/

[pal666]

Yes. My objections to systemd are twofold, and as long as they stand, its excellence in other areas (such as better unit files or daemon management) cannot override. This has been discussed here many times, but heh, recap time :P

1) It does far too much in pid 1. This greatly increases the risk to system stability. "All eggs in one basket" is simply stupid.

2) The attitude and politics of its creators. Subsuming everything like a tentacle monster, actively working against loose coupling.

you confuse facts with your misinformed fantasies
1) systemd does very little as pid 1
2) its creators actively work against harmful linux fragmentation