[svlug] Inits (was: On the process of picking up systemd)

Steve Litt slitt at troubleshooters.com
Sat Jan 3 23:03:05 PST 2015


On Sat, 03 Jan 2015 13:05:52 +0100
Ivan Sergio Borgonovo <mail at webthatworks.it> wrote:

> On 01/03/2015 12:41 AM, Rick Moen wrote:
> 
> > You would have understood this if you had taken my advice and read
> > the 'sysvinit: the eternal red herring' section of
> > http://uselessd.darknedgy.net/ProSystemdAntiSystemd/.  Please do
> > that now, before posting more non-sequitur time-wasting arguments.
> > Thank you.
> 
> I did. It just seems to me that Gentoo, *BSD and Slackware weren't
> deployed at large and that the tools you've mentioned were just
> «research projects and anything [I] see as “toy projects”»

"Research projects" and "toy projects" are meaningless
characterizations.

> 
> But then you may be right, systemd was "forced" on a minority of
> people that "typically hail from more niche distributions like
> Slackware, Gentoo, CRUX and others"...

Almost everyone was forced. Only some, and they tended to be the kind
who would use Slack, Gentoo, etc, minded being forced.

> 
> For their use case systemd is "technically inferior".
> 
> Are you saying that people at RH, Ubuntu, SuSE, Debian (Google??) were
> tricked into choosing a "technically inferior" solution even if they
> knew there were "superior" viable alternatives?

Tricked or pushed, yes.

> Were all their users, including the one that pay their bills and
> companies with large deployments, tricked to accept a "technically
> inferior" solution?

So far, yes. This will likely change.

> Don't you think that in general companies that use Linux even if they
> had enough resources to write their init system would be happier if
> the most adopted init system was one that fit their needs?

Yes, but the most adopted one may *not* fit their needs.

> Don't you think that RH (or the conspirational mansonry behind them)
> would prefer to go with an init system that would make other large
> contributors to the Linux stack happy?

Resounding NO!

> [1]
> 
> I'd say that Slackware, Gentoo, *BSD, CRUX "use case" is technically
> irrelevant and well they are free to support their use case with
> whatever init they prefer as they are doing. (I'm sorry for *BSD but
> well... I'm sorry for many things).
> 
> And don't get tricked by the fact that I tend to stay in group a) for
> Linux system stuff (that generally I consider an appliance).

The words in the preceding parentheses account a lot for our
disagreement. If I considered a computer an appliance, I wouldn't give
a flying flamingo what init system they put on it, and when my
appliance broke I wouldn't give a flying flamingo about which component
broke, I'd just want it fixed. Of course, if *I* were the one who had
to fix it, and the init system went bad, I'd be a lot better off
troubleshooting a system initted with runit or Epoch than systemd.

>  When it
> comes to software development I'm faced everyday with the problem of
> which poison I've to take.
> 
> I do say *shit* all the time about technical choices of other
> developers that substantially enable me to write my software.
> "This web stuff has been written by a million monkeys." (cit.)
> 
> Isn't it frustrating when you ask people to scratch your own itch and
> they keep on missing the "right" point?

Yeah, everyone keeps coming back to that. Nobody wanted Redhat or
Debian to write a wonderful init, or even package a wonderful init. All
we asked was "first, do no harm."


> 
> We're not talking about which DE has to rule the world... seriously...
> what kind of market are Linux DE contending?
> We aren't talking about an unmaintained debugging tool for a niche
> processor (oh shit, you #@!!@).
> We aren't talking about any php[2] module for drupal as well (oh shit
> shit shit and more shit).
> 
> All the stuff they placed in systemd is pretty important where Linux
> is really economically relevant (embedded and cloud?).

A) I'm not convinced the preceding is true, and B) I know for a fact
that many supposed systemd features and benefits could have been kept
in the user space outside of the kernel and init system.

> Were all the players *seriously* involved sleeping under a rock?

Yes they were. Had they not been, a better init system than sysvinit or
Upstart would have been in place by 2010 or before. Had that better
init been in place, Red Hat could not have marketed systemd to anyone
but their downstreams.

I can't respond to the remainder of your email because I didn't
understand it.

SteveT

Steve Litt                *  http://www.troubleshooters.com/
Troubleshooting Training  *  Human Performance




More information about the svlug mailing list