Hello Alex, CC'ing the systemd packager list, if you don't mind. Keeping fullquote for that.
Alex Mayer [2015-06-07 19:34 +0000]: > Hey Martin, > > i think two systemd 218-4 changes in /lib/systemd/system/[email protected] are > the issue. > > > Type=oneshot > When no network available, it causes the service to take 75s. I have removed > this line temporarily. Right, there will be some DHCP timeouts until it fails. However, this rule should only be triggered via udev rules on hotplug events. Does that "take 75s" have any visible effect, i. e. do they stall other units that depend on it which shouldn't be stalled? > ExecStartPost=/sbin/ifquery --state %I > When network available, it lets ifup service fail. (without specified > Type=oneshot) The Type= shouldn't matter here, does it? Can you please be more detailled how it fails exactly, i. e. give me the output of systemctl status -l ifup@*.service and your /etc/network/interfaces? The unit that fails, does that belong to an actual network interface that gets brought up and is online, or something else? > For now i have removed both lines to restore systemd 215-18 like > functionality. > (At least the ifquery call may be removed as soon ifupdown bug #773539 is > fixed.) Right, I really don't like that workaround, but it doesn't seem right to let the unit succeed even if ifup'ing the interface actually fails. Thanks, Martin -- Martin Pitt | http://www.piware.de Ubuntu Developer (www.ubuntu.com) | Debian Developer (www.debian.org) _______________________________________________ Pkg-systemd-maintainers mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-systemd-maintainers
