DNS resolution failures, running ntp module failed, and 401 error during commissioning

[Jason Hobbs]

Early in commissioning, DNS resolution is failing:

For example:

Err:1 http://archive.ubuntu.com/ubuntu xenial InRelease
  Temporary failure resolving 'squid.internal'

and:
015:19:43.163298 * Could not resolve host: entropy.ubuntu.com#01215:19:43.163330 * Closing connection 0#012curl: (6) Could not resolve host: entropy.ubuntu.com

Then we get this warning:
2017-11-07 15:19:46,152 - util.py[WARNING]: Running module ntp (<module 'cloudinit.config.cc_ntp' from '/usr/lib/python3/dist-packages/cloudinit/config/cc_ntp.py'>) failed

Later during commissioning we get 401 errors talking to MAAS:
request to http://10.245.208.33/MAAS/metadata//2012-03-01//maas-scripts/ failed. sleeping 1.: HTTP Error 401: OK

More log: http://paste.ubuntu.com/25911508/

It seems like DNS isn't setup properly at first, but does get setup later. If I tell MAAS to keep the node up after commissioning, I can ssh in and see that resolv.conf is setup.

It looks like DNS isn't working until after the node performs dhcp. This snippet of syslog shows it failing before and working after the dhcp:
http://paste.ubuntu.com/25911535/

I suspect this is related to bug 1711760 and that it isn't properly fixed yet.

This is with 2.3.0~rc2 (6395-g1a078b4-0ubuntu1~16.04.1).