Bug#736258: acpid won't stop, won't upgrade (systemd) - https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=736258

Henrique de Moraes Holschuh hmh at debian.org
Mon Aug 11 00:09:10 BST 2014


On Sun, 10 Aug 2014, Russ Allbery wrote:
> To take a step back, I think this is a fair summary of the discussion.
> Please let me know if you disagree:

...

I agree to all points.  It is a fair summary.

> I think you've convinced me that the approach of stopping a socket by the
> same name when stopping a service in invoke-rc.d stop (which means
> starting it again with invoke-rc.d start) is the safer approach.  At this
> point in the transition and release cycle, I think erring on the side of
> taking the safer approach at the cost of some features is best.  I think
> everyone agrees that there should be some way to control this at the
> invoke-rc.d layer for package maintainers to manipulate.
> 
> This is clearly something that needs some further attention for good
> integration in the longer term.

The "service" utility seems to already have the necessary code to stop both
service and socket at the same time, so we could borrow from that.  It is
important to take a look at the other bug reports mentioned in this bug
report thread, as they have some very useful information (such as the one
about "service" already stopping the socket).

I've already proposed a possible API extension for invoke-rc.d that would
let packages request the "please don't stop the socket" behavior.  If
everybody is ok with the proposal (including the non-imaginative name for
the invoke-rc.d --option :-) ), it is just a matter of coming up with a
patch for invoke-rc.d.

If someone doesn't like something in that API, please reply, or preferably
propose changes to it.  One change I did think about is that we should add a
negative form of the option to the invoke-rc.d API, as that might prove
useful down the road.

I've since *carefully* read all of the systemd documentation for services
and sockets, so I have as much of the full picture as I can have without
actually reading the systemd source code and experimenting with it.  At this
point, I think I could write the patch without doing anything too stupid,
but I will be short on time until I get some other unrelated stuff (but
related to Debian, so it eats into my Debian time) done.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh




More information about the Pkg-systemd-maintainers mailing list