Hi Am 26.07.2014 19:44, schrieb Henrique de Moraes Holschuh: > Since just about everything in a "modern desktop" will pester the acpi > netlink socket, please ask a systemd expert whether "systemctl stop" > *blocks* socket activation of that service until further notice, or not. > If it doesn't block socket activation, then we have a grave problem in the > systemd invoke-rc.d interface. > > The semanthics for invoke-rc.d stop is that the service must NOT be > restarted automatically until it is manually restarted by the admin or by > invoke-rc.d. These semanthics are *required* for proper package > upgrades/downgrades. Note that a direct call to systemctl does count as > "manual invocation by the admin". invoke-rc.d is a (low level) interface which should not be used directly by the administrator and is supposed to be used in the maintainer scripts. invoke-rc.d stop <foo> will only stop the underlying systemd .service, not the socket. A systemctl stop foo.service will only stop the .service file. The "service" command, which is supposed to be used by admins, on the other hand does stop both .service and, if existing, the .socket unit [1]. Up until now, we do not generate explicit maintainer scripts code to stop the socket during upgrades. The underlying idea was, that keeping the socket open would allow to not lose any incoming message. Unfortunately, triggering a service start mid way through the upgrade is not a good idea. We'd need a facility in systemd to defer/block incoming requests until the service is ready. Since we don't have that, we concluded that dh-systemd should generate maintainer scripts code to stop the socket as well as the service. See [1]. Cheers, Michael [1] http://sources.debian.net/src/sysvinit/2.88dsf-53.2/debian/service/service#L191 [2] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=751741 -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
Attachment:
signature.asc
Description: OpenPGP digital signature