Olaf Meeuwissen wrote:

>Installing a desktop, by default, pulls in wicd (or network-manager).
>You can prevent this by using apt-get's --no-install-recommend option.

Not an option within the de??an installer - which was the context of the
original post.


>Whether either package blatantly ignores your static IP configuration
>from when you installed, I cannot tell for sure (zapped wicd) but I
>vaguely remember that you can tell wicd to leave certain interfaces
>unmolested.  That may even be its default behaviour for interfaces that
>are configured in /etc/network/interfaces.
>
[Steve Litt] 
>>I would sure find this behavior surprising.
>
>If wicd breaks static IP address configurations out-of-the-box I'd be
>surprised too.  I've mainly used it in DHCP settings.  On my server's
>wicd was never installed so any static IP configurations just worked as
>intended.

However surprising to any of you, this is my testimony. A statically
configured interface present in /etc/network/interfaces was ignored **as
installed by Devuan ASCII.iso**. Removing wicd fixed the problem. What other
conclusion can reasonably be drawn but that wicd is the one doing the
ignoring?


># Veteran Unix Administrator's are free to cobble together their own
># solution and `apt purge wicd` goes a long ways towards that end ;-P

But that's only an option after the fault, which only shows up after
restarting. If the device is not within easy reach, how will you get a
command line at the random assigned IP address in the first place?

Background: My situation, which you so deride, is a test install of a server
with changing products *and management tools* which is where the desktop
comes in. When a configuration is finally settled upon, the server will be
wiped and reinstalled in a production configuration without desktop as all
my other servers have been, and the management tools installed on a
management workstation. Until then the churn will be limited to this one
system.

In any case, if wicd-vs-static IP is a long-standing issue I agree neither
of us would have encountered it before.

_______________________________________________
Dng mailing list
Dng@lists.dyne.org
https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng

Reply via email to