Ok, I made some progress wrt the original question (still below). I found /etc/maas/dhcpd.conf contained option domain-name-servers 10.0.3.15, and changed it to 192.168.0.11. After restarting the daemon, I now see "node" getting the right DNS, unfortunately this doesn't fix the main problem, which I believe is the reference to 169.254.169.254. It does introduce a new question: while the remaining information from /etc/maas/dhcp.conf is present in the maas GUI, there is no field to enter the dns address. Why? Anyway, my original problem still stands... Any idea?

Original question follows.

In VirtualBox, I have:

• master VM:

  • ubuntu 12.04.3 server
  • eth0: Internal Network, IP= 192.168.0.11
  • eth1: NAT, IP= 10.0.3.15
  • eth2: Host-only, IP= 192.168.56.102
  • running MAAS region and cluster controlller, with DHCP and DNS enabled

• node VM:

  • eth0: Internal Network

node VM boots in PXEboot. DHCP succeeds, and the boot process starts, but during boot I see some issues. One of them is "disk drive not ready yet or not present" for / and /tmp. I've googled this issue, and some people say it happens when the fisical disk is a SSD, which is my case. Anywaythe system seems to recover from this eventually.

Immediately after it starts printing a lot of messages of the form:

2013-10-01 16:52:37,142 - url_helper.py[WARNING]: Calling
'http://169.254.168.254/2009-04-04/meta-data/instance-id failed [x/y]:
 url error [[Errno 113] No route to host]

That IP address is clearly bogous, not sure where it came from. Before that point, I had seen the following network configuration:

  address: 192.168.0.100 
broadcast: 192.168.0.255 
  netmask: 255.255.255.0
  gateway: 192.168.0.1
    dns0 : 10.0.3.15 
    dns1 : 0.0.0.0

Not sure if related, but the dns doesn't seem right, as node doesn't have an interface to reach 10.0.3.15. If that's the problem, what should I change to have the DNS point to 192.168.0.11?

Thanks, Roberto