8:41 a.m.
Hey,
Here's a patch which takes a stab at implementing kickstart's network
command in livecd-creator.
Cheers,
Mark.
12:15 p.m.
On Wed, 2007-04-11 at 14:41 +0100, Mark McLoughlin wrote:
Here's a patch which takes a stab at implementing
kickstart's network
command in livecd-creator.
This ends up changing the behavior, though, in the case that you don't
have a network line at all (which is pretty reasonable to do, especially
if you're setting up for NetworkManager). Also, if you have multiple
network lines and the first specifies the hostname, then it gets
overriden by later ones.
Jeremy
12:29 p.m.
On Wed, 2007-04-11 at 13:15 -0400, Jeremy Katz wrote:
On Wed, 2007-04-11 at 14:41 +0100, Mark McLoughlin wrote:
> Here's a patch which takes a stab at implementing kickstart's network
> command in livecd-creator.
This ends up changing the behavior, though, in the case that you don't
have a network line at all (which is pretty reasonable to do, especially
if you're setting up for NetworkManager). Also, if you have multiple
network lines and the first specifies the hostname, then it gets
overriden by later ones.
So, what you're basically saying is that writeNetworkConfig(),
writeNetworkHosts() and writeNetworkResolv() should not be called per
network line, but rather at the end either with values collected from
the network lines in aggregate, or default values where there are no
network lines?
That makes sense, I'll take another look at anaconda's logic for
this ...
Cheers,
Mark.
12:34 p.m.
On Wed, 2007-04-11 at 18:29 +0100, Mark McLoughlin wrote:
On Wed, 2007-04-11 at 13:15 -0400, Jeremy Katz wrote:
> On Wed, 2007-04-11 at 14:41 +0100, Mark McLoughlin wrote:
> > Here's a patch which takes a stab at implementing kickstart's network
> > command in livecd-creator.
>
> This ends up changing the behavior, though, in the case that you don't
> have a network line at all (which is pretty reasonable to do, especially
> if you're setting up for NetworkManager). Also, if you have multiple
> network lines and the first specifies the hostname, then it gets
> overriden by later ones.
So, what you're basically saying is that writeNetworkConfig(),
writeNetworkHosts() and writeNetworkResolv() should not be called per
network line, but rather at the end either with values collected from
the network lines in aggregate, or default values where there are no
network lines?
Yep. And written that way it makes a whole lot more sense than what I
wrote :-)
Jeremy
4:52 a.m.
On Wed, 2007-04-11 at 18:29 +0100, Mark McLoughlin wrote:
On Wed, 2007-04-11 at 13:15 -0400, Jeremy Katz wrote:
> On Wed, 2007-04-11 at 14:41 +0100, Mark McLoughlin wrote:
> > Here's a patch which takes a stab at implementing kickstart's network
> > command in livecd-creator.
>
> This ends up changing the behavior, though, in the case that you don't
> have a network line at all (which is pretty reasonable to do, especially
> if you're setting up for NetworkManager). Also, if you have multiple
> network lines and the first specifies the hostname, then it gets
> overriden by later ones.
So, what you're basically saying is that writeNetworkConfig(),
writeNetworkHosts() and writeNetworkResolv() should not be called per
network line, but rather at the end either with values collected from
the network lines in aggregate, or default values where there are no
network lines?
This should do it, I think ...
Cheers,
Mark.
3:05 p.m.
On Fri, 2007-04-13 at 10:52 +0100, Mark McLoughlin wrote:
On Wed, 2007-04-11 at 18:29 +0100, Mark McLoughlin wrote:
> On Wed, 2007-04-11 at 13:15 -0400, Jeremy Katz wrote:
> > On Wed, 2007-04-11 at 14:41 +0100, Mark McLoughlin wrote:
> > > Here's a patch which takes a stab at implementing kickstart's
network
> > > command in livecd-creator.
> >
> > This ends up changing the behavior, though, in the case that you don't
> > have a network line at all (which is pretty reasonable to do, especially
> > if you're setting up for NetworkManager). Also, if you have multiple
> > network lines and the first specifies the hostname, then it gets
> > overriden by later ones.
>
> So, what you're basically saying is that writeNetworkConfig(),
> writeNetworkHosts() and writeNetworkResolv() should not be called per
> network line, but rather at the end either with values collected from
> the network lines in aggregate, or default values where there are no
> network lines?
This should do it, I think ...
Looks good. Committed with the tiny fix of using /etc/hosts instead
of /etc/sysconfig hosts.
Jeremy
6220
days inactive
6225
days old
livecd@lists.fedoraproject.org
6 comments
2 participants
participants (2)
-
Jeremy Katz -
Mark McLoughlin