[Install] for static systemd unit file?
biebl at debian.org
Fri Mar 3 21:14:30 GMT 2017
Am 01.03.2017 um 21:51 schrieb Patrick Schleizer:
> Michael Biebl:
>> Am 01.03.2017 um 21:35 schrieb Patrick Schleizer:
>>> How should the [Install] section for static systemd unit file look like?
>> The obvious question is: why does this service need to be statically
> Given the example... With this socket / service file combination, I
> wouldn't know how to enable the service non-statically. In the current
> implementation it looks to me right, and works.
Just wanted to add that this was partially a misunderstanding on my
part. When Patrick was talking about a static service, I was thinking
about one which is enabled by shipping symlinks in the package in /lib/.
I guess what Patrick wants is a service which simply has no [Install]
section at all, because it's not activated by being pulled in via a
target but via some other triggers.
This is perfectly fine, fwiw. Lintian is just overly ambitious here.
Instead of inventing a WantedBy=none.target, it's better to just ignore
or override that lintian warning.
That all said, socket activation is not only about lazy loading, as you
mentioned in a later email. It's more about avoid explicit dependencies.
It's perfectly fine to have a socket-activated service which is also
pulled in by a target like multi-user.target.
Hope that clarifies. Sorry for the confusion this might have caused.
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 833 bytes
Desc: OpenPGP digital signature
More information about the Pkg-systemd-maintainers