Failed to connect to bus: No such file or directory

[rvrignaud]

Hello,

It seems that package behavior changed for 0.12.
I'm not able to install telegraf debian package 0.12.0-1 inside a docker container anymore (on image ubuntu:1604).

Following is not working:

FROM ubuntu:16.04
RUN apt-get update
RUN apt-get install -y wget
RUN wget http://get.influxdb.org/telegraf/telegraf_0.12.0-1_amd64.deb
RUN dpkg -i telegraf_0.12.0-1_amd64.deb

Step 5 : RUN dpkg -i telegraf_0.12.0-1_amd64.deb
 ---> Running in 32380922ea7a
Selecting previously unselected package telegraf.
(Reading database ... 7528 files and directories currently installed.)
Preparing to unpack telegraf_0.12.0-1_amd64.deb ...
Unpacking telegraf (0.12.0-1) ...
Setting up telegraf (0.12.0-1) ...
Created symlink /etc/systemd/system/telegraf.service, pointing to /lib/systemd/system/telegraf.service.
Created symlink /etc/systemd/system/multi-user.target.wants/telegraf.service, pointing to /lib/systemd/system/telegraf.service.
Failed to connect to bus: No such file or directory
Failed to connect to bus: No such file or directory
dpkg: error processing package telegraf (--install):
 subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
 telegraf
The command '/bin/sh -c dpkg -i telegraf_0.12.0-1_amd64.deb' returned a non-zero code: 1

Whereas telegraf 0.11 is working:

FROM ubuntu:16.04
RUN apt-get update
RUN apt-get install -y wget
RUN wget http://get.influxdb.org/telegraf/telegraf_0.11.0-1_amd64.deb
RUN dpkg -i telegraf_0.11.0-1_amd64.deb

[sparrc]

this is the same issue as #1016, it appears that your system's systemd is broken

[rvrignaud]

Hi @sparrc,

My systemd is not broken, it's only not running as I'm in a container build.
It's a perfectly legitimate situation so IMHO there is a problem with the post-install script that require systemd to be running (which is not the case in a docker container). Again the behaviour changed between 0.11 and 0.12.

I think this issue should be reopenned.

[sparrc]

why don't you use the tarballs then? how would you expect to install a debian service without a service manager running?

[rvrignaud]

In case of telegraf, I could indeed use a tarbal, but it's less convenient.
I'm not sure how is managed the service installation but I install a lot of daemon inside a container with debian packages (nginx, bind, postfix, etc ) none of them are failing because systemd is not running.

[sparrc]

@rvrignaud And how do you control those services? do you use a service manager?

[rvrignaud]

This services could be started directly by command / entry point or using a small init system like supervisord.
Any way at docker build time, the init manager will not be running.
Maybe the solution is just not to fail and not exit != 0 in post-install script if systemd is not running ?