On 2015-08-10 at 11:16 +0200, Reindl Harald wrote: > > Am 10.08.2015 um 07:57 schrieb Ivan Shapovalov: > > On 2015-08-06 at 15:01 +0200, Michael Biebl wrote: > > > 2015-08-06 14:43 GMT+02:00 Reindl Harald <h.rei...@thelounge.net> > > > : > > > > well, but "Type=simple" is default and recommended everywhere > > > > because there > > > > is no main-pid to guess and the with "Restart=always" monitored > > > > is > > > > in fact > > > > "ExecStart" > > > > > > Actually, Type=simple has its own share of problems: > > > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=778913 > > > > That's a direct consequence of Type=simple semantics. From > > systemd's > > standpoint, a daemon which crashes 10ms after start-up due to bad > > config is indistinguishable from a daemon which crashes 10hrs after > > starting due to an internal error. > > > > To make things more robust, the daemon should either *properly* > > implement Type=forking (i. e. fork only after initial start-up) > > or it should implement Type=notify > > sorry but in context of the topic that's wrong > > * "RuntimeDirectory" is a service configuration > * the daemon is started as unprivileged user > * "RuntimeDirectory" should be created long before > ExecStart / ExecStartPost and hence i wonder how > there can exist a race condition at all, there > is no valid reason trying to create it twice
I was replying to a specific message. My reply has nothing to do with original problem (which is already fixed in PR 896). -- Ivan Shapovalov / intelfx /
signature.asc
Description: This is a digitally signed message part
_______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
