systemd (Was Re: tmpfs for strategic directories)

Lennart Poettering mzerqung at
Wed May 26 16:14:18 UTC 2010

On Wed, 26.05.10 12:35, Scott James Remnant (scott at wrote:

> > 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).
> > 
> Sorry, but that's complete bullshit.
> We did sit down and discuss things, and you convinced me that
> launchd-style activation was a useful thing to have.  Then you went off
> and wrote systemd anyway.

Well, my reading of this is different. You didn't seem convinced back
then and nothing ever landed in the Upstart tree.

The cgroup start you outright rejected a couple of times since then.

> And it was Ryan Lortie who convinced me that D-Bus was the right way to
> go very early on, the conversion of Upstart to D-Bus happened years ago
> (Fedora is lagging behind on versions so only just got that).

Ok, if you claim that we didn't play a role then I won't repeat that
anymore. ;-)

> > 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.
> > 
> Given that I have changed direction a couple of times with Upstart, and
> have been swayed to different courses by a good argument, I refuse that
> it's "unlikely to change direction" ;-)

Well, so you are willing to get rid of the evnt logic because it is

> I'm certainly more interested in getting Upstart *right* than in rushing
> to get it finished by a certain date.

Hmm, so far the "right" part didn't really work out, did it? ;-)

> To be clear, I believe the reason you implemented systemd instead of
> contributing help to Upstart is:
>  a) your personal distaste for Ubuntu and Canonical

Oh come on. Thanks for turning this into something personal. I mean, I
see how it can be handy to make this personal, because then you can just
say that systemd is born out of a personal vendetta or something and
hence ignore it or so. But uh, this is just stupid, especially since I
have no distaste at all for Ubuntu or Canonical, that is completely made

I mean, I have publicly criticized some of the work Canonical has been
doing in the past on technical grounds, but otoh I have been defending
you too. For example, in this very recent blog story i only said good
things about you:

And also I have been attending a few UDSes, and am working with your
folks on various audio related things. I even hat a longer chat with
your desktop manager at GUADEC last year to improve cooperation. If I
would "dislike" you that much, why would I do this?

There is no distaste. Maybe you wish there was, but that is FUD, und
unfair, and personal accusation. Stop that.

And learn to distuingish "technical criticism" from "emotional

>  b) your personal distaste for the copyright assignment policy (RedHat
>     have signed this agreement for Upstart fwiw)

Oh yes, as if it was only me. I think the majority of free software
folks dislike copyright assignment. 

And I am not even very vocal about all of this, I am not a lawyer, and I
pretty good at staying out of any discussions like that.

Also, rewriting something because of copyright assignment is stupid. As
long as it is Free Software you can simply fork it, nothing more.

I mean, copyright assignment is just annoying for contributors, it gives
potential contributors the clear signal that you are not interested in
their code, and that if you take it you want to benefit from it more
than the contributors could. That's just a dick move.

>  c) your personal love of nih ;-)

Oh, really? this is unsulting. May I quote my blog story:

"Is this an NIH  project?
    Well, I hope that I managed to explain in the text above why we came
    up with something new, instead of building on Upstart or launchd. We
    came up with systemd due to technical reasons, not political

This is just annoying. 

I mean, I read the upstart sources very closely, and I like the code and
publicly stated that. And I have contrubuted code to numerious other
open source projects. And that should tell you that this is not about
NIH, for whatever that actually should be.

> I don't think that's a bad thing, I certainly share (c) in equal
> measures <g>.  I'm also not going to argue that Fedora shouldn't chose a
> different init system to mine, that's not really my place to do so.

Well, nice rethorical trick, but uh, it's just bogus. I have no distaste
for Ubuntu/Canonical, and the copyright assignment mess certainly didn't
help to make me love upstart, but also certainly only played a minor
role in the reasoning for systemd. And the NIH claim is just bogus.

I'd prefer it we would keep this discussion purely technical. Everything
else does not help at all in this matter.

So, if you have say something, then please criticise systemd on a
technical level, and tell us why Upstart is better in your opinion. I am
pretty sure most folks on this list appreciate technical discussions
much more than personal flaming. Thank you.


> However I do dispute that I haven't been flexible wrt Upstart's design;
> indeed I would claim that part of the reason development is slower than
> the rapid against-the-wall pace of other projects, is that I'm too
> flexible with its design ;-)

That doesn't really sound very inviting. I mean "Hey, adopt Upstart now,
we haven't made our mind up yet what we want to be, and we will redesign
the whole story soon!" is not really useful as an advertisement for
Upstart, is it?


Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net           GnuPG 0x1A015CC4

More information about the devel mailing list