[PATCH] Fix debian init script problem for Container based virtualization
Hi,
I just now submitted a debian bug for this because I hope that it gets fixed in Debian stable, but I also wanted to put this on the list (the bug does not have an ID yet, I will put it in the thread). I post this here, because there is an init-script in the xymon-sources debian/ directory (debian/xymon-client.init).
I discovered then when having a debian host where (in my case) LXC virtual machines run, it is impossible to start the xymon-client there, when another xymon-client is already running in one of the containers.
This seems to be, because the invocation of start-stop-daemon does only check the process name.
The problem can be resolved by adding the --pidfile argument to the start-stop-daemon invocations so that start-stop-daemon does not only check the process name but also the PID. During a quick test, the attached patch was fixing the problem.
Kind Regards,
/jbn
-- Joel Brunenberg - Troisdorf PGP: 0xAD25981C
Hi,
debian assigned bug id 819976
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=819976
Regards,
/jbn
-- Joel Brunenberg - Troisdorf PGP: 0xAD25981C
Hi Joel,
On Mon, Apr 04, 2016 at 03:50:04PM +0200, Joel Brunenberg wrote:
I just now submitted a debian bug for this because I hope that it gets fixed in Debian stable, [...] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=819976
Bugs of severity "normal" usually don't get fixed/backported to stable. The risk that the fix breaks something else and makes the situation even worse in stable is too high.
But we will very likely fix this in Unstable/Testing and hence for the next Debian Stable release dubbed "Stretch" (Debian 9).
Kind regards, Axel Beckert
-- Axel Beckert <beckert at phys.ethz.ch> support: +41 44 633 26 68 IT Services Group, HPT H 6 voice: +41 44 633 41 89 Departement of Physics, ETH Zurich CH-8093 Zurich, Switzerland http://nic.phys.ethz.ch/
Hi Axel,
On Thu, Apr 07, 2016 at 12:01:42AM +0200, Axel Beckert wrote:
On Mon, Apr 04, 2016 at 03:50:04PM +0200, Joel Brunenberg wrote:
I just now submitted a debian bug for this because I hope that it gets fixed in Debian stable, [...] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=819976
Bugs of severity "normal" usually don't get fixed/backported to stable. The risk that the fix breaks something else and makes the situation even worse in stable is too high.
I understand. I do not like hearing that very much but it makes sense. Even though this is a tiny incision, I do not feel comfortable projecting all possible effects on the strangest of scenarios this software is run in.
But that being said, the tiny size of this change makes me confident to maintain this change locally for the time being. Effects of this change being removed or overwritten by an update are implicitly monitored really really well. :)
But we will very likely fix this in Unstable/Testing and hence for the next Debian Stable release dubbed "Stretch" (Debian 9).
That is very good news. Another question would be if this is fixed in the package or in the prrovided script in upstream. Even though that does not really matter from an end user perspective, I would be interested to know how that works.
Thanks for the effort,
Joel
-- Joel Brunenberg - Troisdorf PGP: 0xAD25981C
On Mon, Apr 04, 2016 at 03:50:04PM +0200, Joel Brunenberg wrote:
The problem can be resolved by adding the --pidfile argument to the start-stop-daemon invocations so that start-stop-daemon does not only check the process name but also the PID. During a quick test, the attached patch was fixing the problem.
As J.C. Cleaver was pointing out off list, the argument was already present in the "stop" section so adding it there is not necessary. :)
/jbn
-- Joel Brunenberg - Troisdorf PGP: 0xAD25981C
participants (2)
-
beckert@phys.ethz.ch
-
ml@jjim.de