Comment 18 for bug 461725

Revision history for this message
Steve Langasek (vorlon) wrote : Re: [Bug 461725] Re: rc-sysinit job might start before loopback is up

On Wed, Dec 23, 2009 at 02:24:12PM -0000, Martin Pitt wrote:
> What I wonder about is how machines without a lo interface could have
> ever worked sensibly -- you will not even get to gdm without one. Is
> there something else which hardcodes bringing up lo during boot?

So the full story here is that NetworkManager will mask the broken /e/n/i by
helpfully bringing up the lo interface when it starts, *but doesn't call the
ifupdown hooks for it*, so upstart doesn't see that it's up.

> Since that seems to have caused real regressions for apparently quite
> some people (which apparently are not entirely understood yet, i. e. how
> they ended up with a broken interfaces file in the first place), but
> fixed regressions for others (this bug), this seems to be between a rock
> and a hard place, so I wouldn't hectically revert this.

> As a workaround for karmic, could we wait on lo only if "grep -q 'auto
> lo' /etc/network/interfaces"?

Not using upstart's event system.

We should instead arrange to fix /e/n/i on upgrade. I'll use bug #497299 to
track this.

--
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/
<email address hidden> <email address hidden>