[systemd-devel] We are working on Secure Container Applications.
lennart at poettering.net
Mon Jan 9 18:42:19 PST 2012
On Mon, 09.01.12 23:26, "Jóhann B. Guðmundsson" (johannbg at gmail.com) wrote:
> >ReloadRequest: httpd.service
> >Then anyone asking to reload the httpd.service would also cause the
> >httpd_sandbox.service to get the reload.
> Hum should that not happen automatically when you BindTo a unit as
> in if you automatically start/reload/restart/stop an bound unit it
> would also start/reload/restart/stop the unit(s) it's bound to.
No, reload is not propagated like this. The idea is basically that
"bindto" means that under no circumstances the binding unit shall run
without the bound to unit. A corollary of that is that if you stop or
restart the bound to unit (restart just meaning stop followed by start
here), then as a result of the first rule the binding unit must be
restarted too. But for reloading nothing like this really applies.
So yeah, I agree with Dan here, that a new dep type that allows
specifying services to automatically propagate reload requests to makes
a ton of sense.
Lennart Poettering - Red Hat, Inc.
More information about the systemd-devel