On Fri, Jun 15, 2007 at 09:44:53PM -0600, Bob Proulx wrote: > I don't know the full history behind this but looking at the > Bug#300937 I assume the design is that network events are now moved > from boot time action to network event time action to support hotplug > network interfaces. Looking in /usr/share/doc/udev/README.Debian.gz I > see: > > After receiving events about network interfaces, net.agent will call > ifupdown using the --allow=hotplug option. This makes the program act > only on interfaces marked with the "allow-hotplug" statement. > E.g: "allow-hotplug eth0" instead of the usual "auto eth0". > > I have not researched further what it would take to trigger udev to > restart networking but I presume there is a path through it that would > do so. To return to the previous behavior you may change the > allow-hotplug stanzas into auto stanzas and then restarting networking > will restart the network interface. I would welcome further > information on this area. >
It may be that the init.d/networking script needs to be a lot smarter when called with 'restart'. Perhaps it should try to bring all interfaces up that have 'auto' or 'allow-hotplug' but capture the errors generated by non-existing interfaces so that they fail silently. It could even put an info notice in syslog to explain any subsequent error lines there. Or, is there a way for the script to get a list of available (hot-plugged) interfaces from somewhere in /proc and only try to configure ones that exist? Doug. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
