invoke-rc.d gets stuck when trying to start gnukhata-core-engine during post install

[Praveen Arimbrathodiyil]

I'm packaging gnukhata (http://gnukhata.org) for debian and ubuntu but I cannot start the gnukhata-core-engine service during installation on ubuntu. It works fine on debian sid (using systemd). I tried both invoke-rc.d gnukhata-core-engine start as well as service gnukhata-core-engine start. It just go defunct after it actually starts the service. I have tested on Ubuntu 12.04 and 14.04 releases.

You can test it using the packages from http://people.debian.org/~praveen/gnukhata/README (currently postinst doesn not start the service, you can edit debian/gnukhata-core-engine.postinst to start the service and test)

[Praveen Arimbrathodiyil (praveex)]

Blocking itp https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=751673

[Steve Langasek (vorlon)]

And what makes you think that the bug is in a piece of software that's had a stable interface for the past decade and longer, and is used by every Debian service up to and including wheezy, rather than a bug in your invocation?

> Blocking itp https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=751673

That makes no sense at all. Your Debian ITP should not be blocked by an Ubuntu bug. Not that this is an Ubuntu issue anyway.

Closing as invalid. You clearly have something wired up wrong, and there's no reason to think that the problem is in the invoke-rc.d command.

[Praveen Arimbrathodiyil (praveex)]

Steve, that is very helpful way of replying. Now I don't care if it lands in Ubuntu or not. As long as it works on Debian I will be happy. Why did I even took the effort to test it on a Virtual Machine running Ubuntu 14.04!

[Praveen Arimbrathodiyil (praveex)]

It is blocking ITP because I wanted it to run well on Ubuntu and Debian, but not anymore. It is a bug in invoke-rc.d on Ubuntu, because the same command works fine in debian using systemd. If your software is perfect, sorry for opening this bug. I won't touch this again.

[Praveen Arimbrathodiyil (praveex)]

Can someone else look at the package and suggest if we should change the wiring to something else? Consider this as a request for help.

[Praveen Arimbrathodiyil (praveex)]

as per Petter Reinholdtsen's suggestion (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=756114#19), adding db_stop in postinst fixed this problem.