Re: Feedback needed: How to disable services at startup... and keep them so.
On Tue, Dec 7, 2010 at 10:21 AM, Camaleón <noelamac@gmail.com> wrote:
> Hello,
Hi
> A month ago, I disabled Network Manager service in my Squeeze system so
> it doesn't run on start up. I wanted to keep NM installed (just in case)
> but preferred to use the old "ifup" network setup method.
>
> So I issued "update-rc.d network-manager remove" and also disabled gnome
> NM applet from being started. So far so good, no more NM running at
> booting.
Seems like a good way to go.
> A couple of days ago I lost network connectivity in that system (no
> Internet/local access because "eth0" was not getting an IP from dhcp
> server, as used to). I had to restart the whole system in order to
> restore network connectivity (neither restarting networking service nor
> "ifup/down eth0" had any positive effect).
>
> After a bit of digging, I realized that Network Manager service was
> running (!) again. I did not enable, so something happened which caused
> the service to be "reengaged" again.
Cool. Err... I mean, interesting.
> Looking into "/var/log/apt/term.log" I saw the following (sorry, the log
> is recorded in Spanish, hope is still clear):
>
> <snip>
> Basically, the log says on December 1st there was an update for NM and a
> new package got installed (0.8.1-4). After that, even though NM service
> was expressly disabled, the update seems to re-enabled it again.
I don't read spanish, but my french is ok, and tha'ts what I would've
guessed had happened
> Before I fill a bug report (I think a service that has been manually
> disabled should keep its state regardless any further update it can be
> applied afterwards), I would like to get some feedback... what do you
> think on this matter? I missed something -there is a better way to handle
> this or should I write a report?
I'd definitely hold off on the bug report. I think you should look at
the lsb headers of the network-manager script in /etc/init.d and
change them to stop on all levels, start on none, then run "insserv"
without any arguments to again disable network-manager. If this isn't
a future proof method of disabling it then there is definitley a
problem in that packages update maintenance scripts.
Reply to: