[Pkg-sysvinit-devel] Bug#740459: Infinite loop reloading dbus service

[Erich Schubert]

Yes, I'm giving systemd a try; so it makes sense to synchronize state
of insserv and systemd.
I tried upgrading packages such as insserv and systemd; I would need
to check the logs if they were already on experimental before the
problem occurred.
Today I only updated to unstable; but I have upgraded packages to
experimental before.

Maybe insserv is the better place to prevent this from happening than
in the shell script.
The shell script was just easier for me to break the cycle. There may
also be an issue involved with cyclic dependencies in insserv; but I
was unable to pinpoint them. insserv output was completely useless in
trying to figure out where the cycle might be, and the additional
tools (visualizing with dotty etc.) also did not yield anything
corresponding to the error messages. In fact, some of the related
services were missing from the dotty plot; I can only assume that
insserv gets some data from systemd, but the visualization tools do
not.


On Sat, Mar 1, 2014 at 10:04 PM, Petter Reinholdtsen <pere at hungry.com> wrote:
> Control: reassign -1 insserv
> Control: found -1 1.16.0-3
>
> [Erich Schubert]
>> insserv: Note: sysvinit service dbus is shadowed by systemd dbus.service,
>> Forwarding request to '/bin/systemctl --quiet --no-reload enable dbus.service'.
>> Synchronizing state for dbus.service with sysvinit using update-rc.d...
>> Executing /usr/sbin/update-rc.d dbus defaults
>
> This tell me that you are not using the insserv version in unstable,
> but the one in experimental with systemd integration support.  Are you
> using systemd, or only have it installed?
>
>> Please add a safeguard against cyclic invocation of update-rc.d. Thanks.
>
> I suspect this should be fixed in insserv, not in update-rc.d.
> Reassigning there.
>
> --
> Happy hacking
> Petter Reinholdtsen