[systemd-devel] user instance of systemd while inside a container

Keller, Jacob E jacob.e.keller at intel.com
Mon Jul 27 13:55:09 PDT 2015


On Mon, 2015-07-27 at 21:51 +0200, Lennart Poettering wrote:
> On Tue, 14.07.15 18:34, Keller, Jacob E (jacob.e.keller at intel.com) 
> wrote:
> 
> > Hello,
> > 
> > I created an nspawn container which is also running systemd. I 
> > can't
> > figure out why the systemd --user instances aren't started.
> > 
> > I'd like to manage some processes run as a specific user inside the
> > container. Previously I was using a VM to do it this way, and I 
> > would
> > like to be able to port the same code over to the container image.
> > 
> > Is there some setup I need to do to get the instance started? I 
> > tried
> > searching on the web, but came up with very little documentation. 
> > user
> > instances appear to be started by pam.. Maybe pam is not being used 
> > as
> > the login for the container?
> 
> This should just work, and it works for me too, with a Fedora 22 host
> and guest (with current systemd added on top).
> 
> Anything interesting in the logs?
> 
> What does "systemctl status user@$UID.service" say for you after you
> logged in?
> 
> Lennart
> 

Hi, thanks. I found out it was because I am using Centos, which
disables this feature by default. I had to copy a user at .service into
the container, and then things worked fine. Wish downstream projects
wouldn't disable random features like this.... 

Regards,
Jake


More information about the systemd-devel mailing list