[PKG-Openstack-devel] Bug#770706: Bug#770706: Bug#770706: Bug#770706: keystone.service does not start, /var/run/keystone not created

[Gaudenz Steinlin]

Mikaël Cluseau <mcluseau at isi.nc> writes:

> The quick-fix of tmpfiles.d is what I did first and will (probably) work.
>
> But, if I assemble this
>
> On 11/27/2014 11:10 AM, Thomas Goirand wrote:
>>> >Do we agree that with systemd the pid-file logic becomes useless
>> Unfortunately, no, I don't agree. Not with the (poor) way we've designed
>> things in openstack-pkg-tools, which still uses start-stop-daemon.
>
> and this
>
> On 11/27/2014 11:14 AM, Thomas Goirand wrote:
>> Best would be if we could just get rid of the sysv-rc wrapper, and just
>> go the systemd way fully.
>
> I think we can agree that *if* we implement the systemd fully *then* 
> pid-files become useless.
>
> As you say it would be the best way, I would like to (try to) spend a 
> significant time to implement the systemd way this week-end. Just... I 
> don't want to it just for the trash, so I'd like to know if you'd agree 
> with this and if the release team will (probably) accept it, or not.

I agree with the pidfile part. I can't comment on the release team. If
that's important to you best ask them yourself. But even if this does
not end up in jessie we need this for the future. So work in this is
definitely not lost. Maybe you could also implement some of the security
features that systemd provides[1] in the unit files. But that's probably
more than the Release Team is willing to accept.

Gaudenz

[1] http://0pointer.net/public/systemd-nluug-2014.pdf