[systemd-devel] Collection of Unit Files available?

enaut enaut.w at googlemail.com
Thu Dec 2 17:17:10 PST 2010


  On 02.12.2010 23:11, Marius Tolzmann wrote:
> hi again..
>
> On 02.12.2010 22:34, Tom Gundersen wrote:
>>
>> The unit files we use for Arch can be found here:
>> <https://github.com/falconindy/systemd-arch-units>.
>>
>> They have not been tested as much as they should though, so there 
>> might be bugs.
>
> thanks @tom for the link.. it helped..
>
> @all: after inspecting some unit files:
>
> * sshd.socket is marked conflicting with sshd.service
> * sshd.socket implicitly activates sshd at .service since Accept=yes
>   (or Accept=true as mentioned in systemd.socket(5)) is set..
>
> if both (sshd.service and sshd.socket) are enabled which one wins?
>
> i know that this would be kind of a stupid configuration 8) but i 
> think it could be a scenario which may get me more into this 
> dependency thing without studying the sources:
>
> is it sshd.socket because this one is activated in the sockets.target 
> and so before the mutli-user.target where sshd.service would be 
> activated?
>
> or is it sshd.service which triggers stopping sshd.socket because 
> sshd.socket is marked conflicting with it?
>
> and: does it matter where the conflict is defined?
>
> AFAIK if any other unit would require one of those units that unit 
> would "win".. but just in case there is no such later requirement.. ..
>
> marius..
>
> p.s.: here is a link (which was sent to me directly via email) to some 
> units on Lennarts server http://0pointer.de/public/systemd-units/ 
> (just in case anybody else is searching for some example units.. thx 
> to enaut)
doh... Some day I will learn to hit the "reply list" button instead of 
"reply"... But either way I wasn't so sure about sending the archive to 
the list... But either way. If anyone else is interested in the 
exherbo-unit-archive just request it.

Well I just tried it and the sockets are started first and remain 
active. Guess it is because the sockets are started very early and 
systemd has a lazy starting policy (meaning only start what it really 
has to). But starting sshd.service manually stops the sshd.socket.

enaut

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/systemd-devel/attachments/20101203/332348c6/attachment.html>


More information about the systemd-devel mailing list