Am 13.04.2013 00:33, schrieb Michael Biebl: > Looking again at the patches and the changes in the jessie-pending > branch, I think there is one case we are currently not handling > properly, and i.e. disabled services. > > Disabled services should not be started after a package upgrade, as that > would imho be unexpected behaviour and a deviation from what happens > under sysvinit. > > There are basically two variants: > a/ services which use --restart-after-upgrade > b/ the default with uses stop in prerm and start in postinst > > for a/ we should probably use try-restart instead of restart for > disabled services. > > for b/ the current behaviour in SysV iirc is, that the service is > stopped in prerm but *not* started in postinst. I will verify that tomorrow.
I have to add here, that this is in practice really a problem only for systemd-only services, given that the SysV and systemd enable-state are in sync (which will be the case once sECuREs patches for update-rc.d and systemctl are uploaded). invoke-rc.d tests the enabled-state of the SysV init script and will skip the (re)start action, if the service is disabled. That said, I think it wouldn't hurt to explicitly check the state of the systemd service (too), before trying to start it. What do you think? Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
signature.asc
Description: OpenPGP digital signature