On Tue, 2011-07-26 at 23:46 +0200, Lennart Poettering wrote: > On Tue, 26.07.11 17:16, James Antill (james@xxxxxxxxxxxxxxxxx) wrote: > > > > Perhaps it's because in %postun it has "systemctl try-restart > > > sshd.service" where it used to have "service sshd condrestart", and > > > systemd kills all running sshd processes whereas the old one only > > > killed the main one and not the children? > > > > It looks more like the first victim of systemd using cgroups as a > > replacement for setsid() etc. ... pam_systemd being the first return > > volley in that war. *sigh*. > > systemd is not using cgroups as replacement for setsid(). And then you go on to explain how pam_systemd is required so that it can do the setsid() like "remove me from the session" call ... and thus. not get hit with the kill() sent to the previous session. I'm not saying that cgroups doesn't do _other_ things as well ... but using it as a shiner setsid() is just doomed to a war where everyone puts "cgroup calls" where setsid() calls currently are. -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel
