Seems I'm reproducing a milder version of bug #862657.
In my case the only issue I'm seeing is when taking down a bridge, an
interface attached to the bridge is also shut down even though I was
expecting ifupdown to control the interface. Thankfully the damage is
milder than in Sebastian Stellingwerff's case.
The experimentation I've been doing seems to lead to a better "standard"
for how to setup the interfaces file when using bridges (for an interface
br0 bridging eth0 and eth0):
iface eth0 inet manual
# extra commands here, if needed
iface eth1 inet manual
# extra commands here, if needed
iface br0 inet dhcp
bridge_ports eth0 eth1
# as usual for a bridge
Mainly, interfaces marked "manual" may have almost zero commands run to
bring them up/down. In a case like this doing `ifup eth0` will simply
mark the interface as being up and *nothing* else. In which case bridge
utils could simply run `ifup eth0 eth1` and not deal with whether the
child interfaces needed any configuration or not.
This seems a far better solution since ifupdown is already prepared to
deal with both simple interface configurations and complex interface
configurations. Bridge utils is turning into a buggy mess due to trying
to deal with too much internally.
--
(\___(\___(\______ --=> 8-) EHM <=-- ______/)___/)___/)
\BS ( | ehem+sig...@m5p.com PGP 87145445 | ) /
\_CS\ | _____ -O #include <stddisclaimer.h> O- _____ | / _/
8A19\___\_|_/58D2 7E3D DDF4 7BA6 <-PGP-> 41D1 B375 37D0 8714\_|_/___/5445
