Hi Martin, On Fr, 2015-10-23 at 17:10 +0200, Martin Pitt wrote: (..) > > 227-2, sid: Fails, retval 6: > > # root? systemctl restart non-existing.service > > Failed to restart non-existing.service: Unit non-existing.service failed > > to load: No such file or directory. > > # root? systemctl restart non-existing.service > > Failed to restart non-existing.service: Unit non-existing.service failed > > to load: No such file or directory. > > That looks right -- systemctl is supposed to fail for a nonexisting > script. I don't even consider that a bug, but a fix.
hmm, sorry, seems I did a cut-and-paste error here; the last two lines should have been root@manwe:~# systemctl restart wicd.service Failed to restart wicd.service: Unit wicd.service failed to load: No such file or directory. So yes, for a non-existing script, that's (changed behaviour) but correct. For the example wicd service however, it should ignore the call like before, afaiu. (...) > Starting/stopping services in a schroot has never been defined > behaviour, as in general you can't do that. chroots should have a > policy-rc.d (see /usr/share/doc/sysv-rc/README.policy-rc.d.gz) which > disables service starting/stopping from package maintainer scripts. Ahh, ok. Will have a look at that. > Also, package maintainer scripts certainly shouldn't call invoke-rc.d > on a nonexisting service? Well, the "nonexisting" example should only serve to simply demonstrate that systemd actually _has_ changed behaviour. I am concerned about actual packages failing to install or remove in a plain chroot. Hth! S