systemd (Was Re: tmpfs for strategic directories)

[Lennart Poettering]

On Mon, 24.05.10 21:17, Jon Masters (jonathan at jonmasters.org) wrote:

> > > So do you plan on getting it by default for Fedora 14?   :-)
> > 
> > Well, I am certainly planning to have a package for it in F14. But
> > whether we can have it as default is to be seen. 
> 
> This is not intended to start a flamewar.
> 
> I hadn't really heard about systemd, so I did some poking through your
> blog (http://0pointer.de/blog/projects/systemd.html ), which has a very
> lengthy amount on the subject, and thanks for writing that. There's even
> a section entitled "On upstart", which points out a few issues with
> Upstart and other things, such as SMF. But I'm still not really sure how
> to answer the first thing I thought of, which was along the lines of
> "why replace init yet again with something else when Upstart is just now
> getting some traction in the wider community?". 

Well, it would be much worse if we did this later on. The earlier the
better.

> Can you point us to where any background discussion has taken place
> with Upstart folks?

No, I cannot. Kay and I and a couple of others sat down at various LPC
and GUADEC and discussed what we would like to see in an init
system. And we had long discussions, but ultimately most of our ideas
were outright rejected by Scott, such as the launchd-style activation
and the cgroup stuff, two of the most awesome features in systemd
now. (That said, we actually managed to convince him on other points,
i.e. I believe we played a role in turning him from a D-Bus-hater into a
D-Bus-lover).

But anyway, these discussions did happen, over years. But there is no
recorded archive of that, no mailing list discussion I could point you
to, sorry. You can ask Kay and me and Scott about it though.

So we have discussed this with Scott in much detail, and we have
followed his development for a longer time. But in the end I just don't
think Upstart is the right thing, and fundamentally flawed and unlikely
to change direction, which is why we chose to start anew, and not just
"fix" Upstart. For more about that just read my blog story.

In summary, this is not coming out of the blue for Scott, we have
discussed that in length with him, but at one point we just saw that
Upstart is fundamentally not going to be what we have in mind, and so we
started a new experiment, and so far it turned out to have very
convincing results, if I may say in my humbleness.

> The thing is, I like some of the things Fedora re-invents differently.
> For example, dracut is quite nice (though there are too many initramfs
> creation tools available at this point, but then they do tend to be more
> distro specific), but IMO it is better to avoid redoing something at all
> costs unless driven by necessity of innovation. So assuming there's a
> nice backstory here, a need for Fedora to do something different, etc.
> then cool. But if there isn't, I'm much more weary of this.

Well, just read the core design and the feature list on our blog story,
and compare that with Upstart. And I am quite sure that if you grok the
awesomeness of the socket based activation that then you will no longer
wonder why systemd is the future and Upstart is not.

Lennart

-- 
Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net
http://0pointer.net/lennart/           GnuPG 0x1A015CC4