[Some rampant snipping, I'm afraid. Hope that is ok.]

On Thu 01 Feb 2018 at 09:41:36 -0600, David Wright wrote:

> On Thu 01 Feb 2018 at 10:55:35 (+0000), Brian wrote:
> > > 
> > > Intended to do what?
> > 
> > To leave the user without network connectivity after first boot? There
> > are at least three bug reports against netcfg on the matter. My
> > recollection is that no deeper intention is revealed there.
> 
> The two that seem most relevant are #694068 and #709017. Finding out
> that a udeb package called netcfg even exists does of course require
> examination of the logs and disection of the debian-installer.
> 
> The considered opinion of Sorina - Gabriela Sandu <sandu.sor...@gmail.com> is
> 
>      Date: Sun, 25 Nov 2012 17:30:28 +0200
>     "For most cases, I think not adding configuration for wireless in /e/n/i 
> is good"
> 
> which shows a lack of attention to important details like leaving the
> new installation with some sort of connectivity.

No reason given for the goodness is given, which makes it hard to judge.
 
> The considered opinion of Colin Watson <cjwat...@debian.org> Date:
>      Thu, 28 Aug 2014 20:03:55 +0100
> 
>     "This is wrong, I'm afraid.  /etc/network/interfaces will *always*
>      exist at this stage, because it's copied by netcfg's
>      base-installer hook.  However, the finish-install hook is
>      explicitly using "netcfg write_loopback" in some cases to make
>      sure that /etc/network/interfaces contains only the loopback
>      entry.  Declining to copy this to the target system would break
>      those cases.
> 
>     "What I'll do instead is copy /etc/network/interfaces only for the
>      netcfg/target_network_config settings that require it.  Then this
>      may just work for you if you don't have network-manager
>      installed,"
> 
> This implies that "other cases" have been considered but forgotten.
> 
> Yes, I don't think the intentions are deeper, but just that simple
> cases have been overlooked, and overlooked for several years.

This statement is from #709017, a closed report. Fixed with

   * Don't copy /etc/network/interfaces to /target if
     netcfg/target_network_config=ifupdown;

> > How connectivity is re-established on a machine with only a wireless
> > interface is left as an exercise for the user.
> 
> This is some sort of sick joke.

Re-writing /e/n/i again for a wireless interface is expected.

> > You think it is required. I think it is required. Other users think it
> > is required. Perhaps the installer knows best? :)
> 
> Obviously not. It appears that the debian-installer team isn't able to
> summon up the energy to debug one of their scripts and, because they
> never meet the users who run into the bug, they feel no pressure to fix it.
> 
> > It does, however, give
> > a way to go against its wishes:
> > 
> > Template: netcfg/target_network_config
> > Type: select
> > Choices-C: nm_config, ifupdown, loopback
> > Choices: Network Manager, ifupdown (/etc/network/interfaces), No network 
> > configuration
> > Description: for internal use; can be preseeded
> >  Specifies what kind of network connection management tool should be
> >  configured post-installation if multiple are available. Automatic
> >  selection is used in this order when not specified: network-manager if
> >  available (on Linux only), ethernet configuration through ifupdown on wired
> >  installation and loopback configuration through ifupdown on wireless
> >  installations.
> 
> If, by that, you mean preseeding, that's just a crummy workaround for
> not fixing the problem. Preseeding wasn't designed for that.

I'm not very sure I entirely agree with what you write but will not
argue the point. What I would say is that most users' thoughts would
not turn to preseeding as a solution, or leap to writing an /e/n/i if
they meet this issue.
 
> > No DE installed; therefore no network-manager. Installation was not over
> > a wired connection; therefore no ifupdown stanza in /e/n/i.
> 
> I'm being sold a bill of goods here, then, as it's not clear how the
> system can perform as a server if it has no connectivity.
> 
> So the final word is Sorina's:
> "For most cases, I think not adding configuration for wireless in /e/n/i is 
> good"
> 
> Considering events of the past year, I guess that's what now counts as
> policy. It just took the world a while to catch up with Debian.

I've written enough on this over the past five years, both in -user and
elsewhere. I can cope with the situation myself but dislike the jumping
through hoops aspect to get a simple thing like connectivity in d-i.
I've come to the conclusion that the present arrangement is as good as
its going to get. It's difficult to know where to go from here but
discussion on -user is unlikely to lead to any change.

Anyway, you jogged my memory:

  https://lists.debian.org/debian-boot/2012/09/msg00252.html

Sorina - Gabriela Sandu:

 > > I realise a default is only a default and the selection can be changed,
 > > but I'm puzzled by the third option. Why treat a wireless install
 > > differently from a wired install? It would expected that a user who has
 > > chosen not to use a wired connection would still want connectivity after
 > > booting into into the new system,

 > The main reason for this is that as far as I know writing configs
 > related to a wireless network to /e/n/i enforce using only that
 > particular network later (of course if you don't modify the file) and
 > also that the interface is unmanageable for other tools. The idea was
 > to leave the network unconfigured, so that it can be managed later
 > (perhaps via something else than NM).


Gaudenz Steinlin:

 > The main reason for the question was to make it preseedable. So for
 > example automated desktop installs could still choose ifupdown. Or if
 > you know that your configuration management is going to install
 > network-manager afterwards you can choose the network-manager
 > configuration depsite it nm not being installed at the moment. Or you
 > can choose to not configure anything because you have a complex post
 > install script doing network configuration.

 > IIRC we decided on this default before we added the code to change the
 > access mode of /e/n/i if it contains a password. The main reason for
 > defaulting to no configuration in this case was to avoid having
 > passwords in there. If people think it should default to ifupdown in
 > this case this can be changed.

I imagine my "...not aware of any comprehensive explanation..." should
be modified.

-- 
Brian.

Reply via email to