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

Bug#727708: upstart and upgrading from sysvinit scripts



On Sun, Dec 29, 2013 at 01:43:59PM -0800, Nikolaus Rath wrote:
> I'm a bit surprised that you mention this only now, after Russ'
> extensive mail. Could you tell us if there are there other components in
> systemd that you think are similarly flawed,

Why should it have been mentioned before now?  I don't consider socket-based
activation to be a decisive difference between the two systems.  The systemd
implementation of socket-based activation is clearly more featureful than
the upstart implementation, and it sounds like it's also more robust - at
the cost of a significant amount of implementation complexity, and quite a
bit of hidden policy that's difficult to discern.

The fundamental reason for systemd's more complete implementation of
socket-based activation is relative priorities, not that the upstart
developers somehow tried and failed to match systemd.  With an
already-debugged set of dependency declarations, as we have in Debian for
sysvinit/insserv and in Ubuntu for upstart, socket-based activation provides
only incremental benefits which I am skeptical that it's worth Debian
investing in.

I think there's some serious misestimation here of the relative costs to
Debian of resolving any issues in upstart that the members of the TC might
consider blockers, vs. the work to integrate systemd throughout the
distribution - an integration that would need to be evolved from first
principles in a Debian context, not copied wholesale from another
distribution like Fedora with different policies and architectural choices.

To be clear: speaking as part of the upstart upstream team, if the TC
decided that they considered socket-based activation the correct model for
Debian to adopt going forward, and that having a full-featured socket
activation implementation was therefore a precondition for adopting upstart,
I think we would have no trouble at all delivering that for jessie.  There's
already agreement in principle with improving upstart's socket-based
activation support for compatibility with systemd's; and the lack of ipv6
support is something I reported a bug about quite a while ago myself
(https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/942955).  It's only
a question of priorities - improved socket activation support hasn't been a
priority, because the existing facilities have been giving upstart's
downstreams what they need.

> and/or are if there other features in upstart that you think will never
> deliver the benefits one would naively expect from them?

Socket-based activation has never been a feature that upstart upstream has
promoted the use of.

-- 
Steve Langasek                   Give me a lever long enough and a Free OS
Debian Developer                   to set it on, and I can move the world.
Ubuntu Developer                                    http://www.debian.org/
slangasek@ubuntu.com                                     vorlon@debian.org

Attachment: signature.asc
Description: Digital signature


Reply to: