On Di, 22.12.20 15:24, freedesktop@xxxxxxxxxxxxx (freedesktop@xxxxxxxxxxxxx) wrote: 65;6200;1c > The third.service started as soon as *either* the first.target or > second.target was started, even though it has *both* of them as Requisites, > and as previously quoted, the manual, > https://www.freedesktop.org/software/systemd/man/systemd.unit.html#Requisite > =, says: > | Requisite= > | > | Similar to Requires=. However, if the units listed here are not > | started already, they will not be started and the starting of > | this unit will fail immediately. > > Bluntly, that does not appear to be true. It is a documentation bug, or at > least poorly worded. Bluntly, maybe actually read what the docs say. Exactly one sentence after the one you quoted you find this: Requisite= does not imply an ordering dependency, even if both units are started in the same transaction. Hence this setting should usually be combined with After=, to ensure this unit is not started before the other unit. > If first.target is started (which has Wants=third.service), but > second.target is not, then how is it that third.service's > "Requisite=second.target" does not cause "the starting of third.service to > fail immediately"? > > Following Andrei's suggestion, I changed third.service to: Maybe just read the docs, more then two sentences. I mean, the docs certainly can use improvement. Everyhing can. But humm, the docs don't need "fixing", they aren't broken. They suggest exactly what to do. Happy to review/merge a patch that improves things, but I think the brokeness you two imply is just not there. Lennart -- Lennart Poettering, Berlin _______________________________________________ systemd-devel mailing list systemd-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/systemd-devel