[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Switching the default startup method

On Mon, 24 Aug 2009, Bernd Zeimetz wrote:
> No, decimal numbers are the superior design here as they just work without any
> magic. One of the reasons I migrated away from SuSE long time ago was the mess
> called insserv...

There are reports of init script not working because they rely on some
other (unexpected) initialization done by other scripts, i.e. numbers
were wrong and could not be easily fixed. And there's nothing magic in the
dependency based system.

On Mon, 24 Aug 2009, Wouter Verhelst wrote:
> On Mon, Aug 24, 2009 at 08:54:06AM +0200, Raphael Hertzog wrote:
> > On Mon, 24 Aug 2009, Andreas Barth wrote:
> > > We should definitly continue to support oldstyle booting, at least for
> > > the time being.
> > 
> > Until what? Missing boot-time dependencies were the only problem that had
> > to be adressed to fix boot sequence ordering.
> Until we're 100% confident that the new method is working correctly, and
> is not causing problems for our users.

How do you get that confidence without testing it on a large scale such as
unstable? It's not like he did not work on this before-hand. It has been
tested but there's no way he could have tested all combinations.

Furthermore all reports pointed by Andreas concerns the fact that
sysv-rc/insserv can't be easily removed/replaced, all reports about
actual boot failures have been worked on (at least I saw some updates made
by Petter to fix some of them).

> I have no problem with making it the default for new installs -- that
> makes sense. However, it should still be possible for people to remove
> it if they have problems with it.

I don't see the value in being able to disable insserv for those using
sysv-rc. I understand that Petter might want to maintain only a single boot

I agree however that both should be removable together for those who want
to use something else.

> > They are relying on an inferior system and the fact that they are used to
> > it doesn't change anything on its inferior design.
> What makes it an 'inferior design'? I utterly disagree with that statement.

Getting the numbers right is a difficult task and as soon as you want to
order something between 2 services registered with the same priority you
might have to renumber lots of services since you don't know whether one
service was started after another on purpose or not.

With dependency based ordering, you just state the dependencies and you
let it figure out the order.

> There are advantages to dependency-based boot systems, sure; but there
> are advantages to *not* having that, too (e.g., it is more
> deterministic, and therefore easier to debug).

Well, both are deterministic but they do not decide of the ordering in the
same way and it's just easier for our brains to represent a number-based

> Debian's had multiple choices for init scripts for a long time (file-rc
> vs sysv-rc). I don't think there's any good reason to throw that out the
> window.

I'm not asking for that. Neither does Petter:

(Although that possibility of choice might disappear once we switch to
a completely different system like upstart)

On Mon, 24 Aug 2009, Andreas Barth wrote:
> * Raphael Hertzog (hertzog@debian.org) [090824 08:54]:
> > On Mon, 24 Aug 2009, Andreas Barth wrote:
> > > We should definitly continue to support oldstyle booting, at least for
> > > the time being.
> > 
> > Until what?
> Until we know that the new method really works 100% correct, people
> enjoy the switch and we noticed that in fact the old packages are
> orphaned and not used anymore.

In what ways do the bug reports that you pointed out indicate that
the new method doesn't work?

> This is not the case. I don't know when this will be the case, but
> don't expect it to be the case during the squeeze release cycle.
> (1-2 stable releases I'd expect if everything works well.)

Why do you expect such a long time to be needed? The work has been
on-going for a full release cycle already.

> > They are relying on an inferior system and the fact that they are used to
> > it doesn't change anything on its inferior design.
> This is just your own personal opinion. Please do not try to enforce
> your personal opinion on all of us. Thanks.

Why do you believe that Petter's work will improve the distribution if
you don't consider dependency based boot sequence ordering to be superior?

What gains do you expect from his work?

> > That's granted but it's easier to say from your place instead of petter's
> > place... I for one appreciate the work that he has put in all this and
> > I would highly prefer that you help him instead of complaining about his
> > work.
> My proposal is to just undo the dependency for the moment. This would
> take the pressure away.

Or simply merge insserv in sysv-rc. The only problems that people seem to
have are related to the fact that insserv can't be uninstalled and the
reasons why that fails come from the fact that the insserv package has
been designed to be optional so that it could be quickly installed and
removed to help test it.

If merged, the new update-rc.d implementation would be default and other
packages could again divert update-rc.d.

> > Because you're giving away the message that you don't care very much of
> > Petter's work and that you prefer staying with the old system instead of
> > fixing the new system to suit your needs, and that's backwards.
> Can we please stay away from ad-hominem arguments.

Sorry, I did not want to attack you, I just wanted to share my feeling of
how I would interpret your message if I were Petter.

> I however do not appreciate breaking working systems, that is correct.

So please point us to bugs related to breakages on upgrades (there have
been some I know, but I think Petter dealt with them correctly) if you
want to use that argument to not switch to insserv by default. The
current bugs that you pointed out only have to do with file-rc
users that are not happy.

I would like to mention the fact that the new file-rc maintainer is not
really cooperative either (thus not improving the situation for its
users). It would be nice if Alexander pointed out why he doesn't want to
fix 539609, his angry reply in
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=539591#15 doesn't bring
the discussion forward while Petter tried to lay the path to allow file-rc
to be a working alternative again.

> (In other words, not all Debian systems are alike. That Debian allows
> the variaty and we don't say "uh, but all people use
> (gnome|kde|whatever), this is not supported" is one of our (almost
> unique) selling points. For what reason should we do it different this
> time, especially without being forced?)

You can request choice, but the choice should not come at the expense
of not improving our default boot system. 

On Mon, 24 Aug 2009, Philipp Kern wrote:
> "I appreciate that you're working on improving the experience of our
> users during startup, e.g. by adding dependency information to the
> init scripts. I think that will in the long run be good for Debians
> users."

I've read that. But now that all scripts have dependency information
what else shall we wait before benefiting from the good? In other words,
while this has been written full of good intent to avoid upsetting Petter,
it sounds wrong in my head when at the same time he requests to not go
further and to wait an indefinite amount of time.

> We agree that it's nice what he's doing, we disagree about the sudden
> breaking of other packages without prior consensus.  And adding packages to
> the (quasi-)essential set without prior consultation is wrong, too.
> Think how it was discussed to switch to dash and how to do it properly.

It think Petter communicated largely on his goal for quite a long time.
It's been a release goal for Lenny already. He organized a BoF at debconf

Very similar to the dash RG except that he did not say on -devel that he
will do it soon.

(For the sudden breaking of other packages, see my reply to Andreas above)

Raphaël Hertzog

Reply to: