network-manager 0.8 needs to kill 0.7 nm-system-settings daemon in init.d script

[Tony Espy]

Binary package hint: network-manager

The init.d script for NM 0.8 ( /etc/init.d/NetworkManager ) needs to kill a running NM 0.7 nm-system-settings daemon on a restart, otherwise network-manager won't start.

[Nikola Yanev (gericom)]

confirming here too! now i'm using net with pppoeconf settings, coz i can't make NM start !?!
if someone can tell me how to, it gonna be great

using karmic alpha 3 with all current updates
network-manager_0.8~a~git.20090804t185522.4bab334-0ubuntu1
network-manager-gnome_0.8~a~git.20090805t131328.d1edfce-0ubuntu1

[Nikola Yanev (gericom)]

let all of the team be notified ;)

[Nikola Yanev (gericom)]

Hello, anybody? doesn't anybody have this problem?
I did purge to NM reboot restored all files modified by pppoeconf, rebooted again, installed NM, rebooted, and nothing happened, even doesn't allow me to create system connection, an when i execute sudo /etc/init.d/NetworkManager restart it starts connecting to ifupdown (eth0) ??? even it's set to false?
Any ideas?
Thnx

[Tony Espy (awe)]

Nikola, please enter a new bug for the pppoe settings, it's really a separate problem.

The best way to do this is to run the following command from a terminal:

% ubuntu-bug network-manager

I'll try and see if I can re-create this tomorrow.

[Tony Espy (awe)]

I'm not quite sure how to fix this bug.

The original description says that the init.d script should be changed to kill nm-system-settings, however as the init.d script is pre-existing ( or auto-generated ), I'm not sure how to patch it?

Any suggestions would be much appreciated.

Would it suffice to add a killall in network-manager.postinst? I suppose that has the potential to kill an existing systems connection?

[Tony Espy (awe)]

Since NM now uses upstart, the original patch I sent upstream for the Debian init script no longer really applies.

So, I think this should get handled in the postinst script...

Any objections?

[Alexander Sack (asac)]

i think we should look into upstreaming the upstream init script.

[Alexander Sack (asac)]

the _upstart_ init script i ment ;)

[Bryan Quigley (bryanquigley)]

We're moving from upstart now, so this bug doesn't need to be done anymore.