Systemd 217: Many New Features, Even More Bug-Fixes

[samdraz]

Documentation

is there documentation for these features?

[uid313]

systemd have no clearly defined scope.

I think the uselessd (despite its name) actually seems like an interesting project and really promising.

[pal666]

systemd have no clearly defined scope.

I think the uselessd (despite its name) actually seems like an interesting project and really promising.

you have no brains.
systemd's scope is basesystem

[Karl Napf]

systemd have no clearly defined scope.

I think the uselessd (despite its name) actually seems like an interesting project and really promising.

The value in systemd is not primary the init system, but the base infrastructure it provides in the form of daemons run by that init system. Logind, networkd, soon consoled and all the other stuff. That is what other projects depend on.

Uselessd unfortunately removes exactly those parts. So I find it rather useless.

[uid313]

The value in systemd is not primary the init system, but the base infrastructure it provides in the form of daemons run by that init system. Logind, networkd, soon consoled and all the other stuff. That is what other projects depend on.

Uselessd unfortunately removes exactly those parts. So I find it rather useless.

Well, if logind, networkd, and consoled are so good, can't they be ported to run on other init systems?

Because with systemd, you're buying into a huge monolith.

It feels systemd and its developers cannot be trusted, they play politics and ignore the community.

Linus said it, Poettering breaks stuff, writes shitty code, and refuses to fix it, then blames others for it.

[gotwig]

What about dropping the whole coredump concept?

[finalzone]

Well, if logind, networkd, and consoled are so good, can't they be ported to run on other init systems?

You talk about systemd as system management. The init part does not need above optional components should be distribution want.

Because with systemd, you're buying into a huge monolith.
It feels systemd and its developers cannot be trusted, they play politics and ignore the community.

Debilerate misinformation not worth to correct.

Linus said it, Poettering breaks stuff, writes shitty code, and refuses to fix it, then blames others for it.

You are accusing the wrong person.

[Karl Napf]

Well, if logind, networkd, and consoled are so good, can't they be ported to run on other init systems?

In theory you could, but in practice most of these make heavy use of cgroups in the extended systemd flavor. So apparently you need to reimplement the DBus interfaces. Logind-shim and systembsd do exactly that.

More effort like those is needed, if non-systemd init systems are supposed to stay relevant.

Or better yet: Come up with an init system with a similarly powerful cgroup interface, so that those daemons are easy to port to that shiny new init system.

Because with systemd, you're buying into a huge monolith.

Not a monolith, just one consistent system. That is something Linux dearly needed for over a decade now!

It feels systemd and its developers cannot be trusted, they play politics and ignore the community.

All 500 of them are not trustworthy?

If you think politics have no part in open source then you never managed a reasonably sized project. I tell you a secret: Where there are people there is politics.

Which community are they ignoring? All my questions to the systemd mailing list got answered and I even managed to submit a patch which was reviewed and merged without fuss. I definitely do not feel ignored. I guess they are indeed ignoring a lot of the angry muttering, but quite frankly that is the best they can do with that.

Linus said it, Poettering breaks stuff, writes shitty code, and refuses to fix it, then blames others for it.

He was bitching about Kay, not Lennart. And Linus has bitched about all the major kernel developers at some point in time, too. I fail to see how that is relevant.

Finally Linus also said that he had no beef with systems itself in a recent interview.

Just wait a bit and this whole thing will wind down again. Believe me, it was the same when this overly complex sysv init abnomination was introduced to replace that nice and simple BSD style init script we had before it.

[TheBlackCat]

Well, if logind, networkd, and consoled are so good, can't they be ported to run on other init systems?

Sure, if anybody was willing to make another init system with the features they depend on. The problem is that the anti-systemd crowd has convinced themselves that nobody wants those features (despite all the projects using them), so they have made no alternative init system that provides them.

It feels systemd and its developers cannot be trusted, they play politics and ignore the community.

They ignore a tiny group of extremely vocal people, as every software project must. No matter what anyone does with any software, there will always be people who disagree. systemd developers focus their attention on providing the most benefit to the largest number of users, as all software projects must do. This is demonstrated by the fact that it has been adopted by so many projects and so many distros, even including direct competitors to Red Hat and projects with no connection at all to Red Hat, and even including projects and distros that had or have no existing dependency on any project that joined the systemd umbrella.

Linus said it, Poettering breaks stuff, writes shitty code, and refuses to fix it, then blames others for it.

Please provide the exact quote where Linus said this. You can't because he never said it.

[uid313]

Please provide the exact quote where Linus said this. You can't because he never said it.

Sorry, I was mistaken. It wasn't directed to Lennart, but his fellow systemd buddy Kay.

http://www.phoronix.com/scan.php?page=news_item&px=MTY1MzA