Comment 75 for bug 1403152

This kernel bug, is affecting all the container solutions also on Xenial / 16.04.1 up to 4.4.0-38.

Lots of people are hitting this on many distributions / container solutions:
https://github.com/docker/docker/issues/5618 (mentions Docker, LXC, kubernetes, etc.)
https://github.com/coreos/bugs/issues/254
...

Apparently this was fixed in 12.04 and now reintroduced in 16.04.

Once you get the following dmesg message:
unregister_netdevice: waiting for lo to become free. Usage count = 1
... The only fix is to reboot the host.

What I gathered from the docker issue tracking this, it seems a fix was merged in net-next very recently:
http://www.spinics.net/lists/netdev/msg393688.html

... and is in mainline now:
https://github.com/torvalds/linux/blob/master/net/ipv6/addrconf.c

As I wasn't sure this thread was tracking Xenial (I just added it to the list), I'm trying to figure out if Ubuntu is tracking this issue upstream and can backport it into 16.04's 4.4.0-? branch, as a lot of container use case start to break down because of this bug for the past weeks?

Thanks!