IBM ppc64le MAAS images booting with no network

[Mike Rushton]

The latest daily Xenial image for MAAS seems to lack network connectivity. Tested with kernel version 4.4.0-15 from the 20160328 image from the daily stream

Attached are logs from the boot sequence as well as the maas deployment logs.

[Brad Figg (brad-figg)]

This bug is missing log files that will aid in diagnosing the problem. From a terminal window please run:

apport-collect 1563055

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Mike Rushton (leftyfb)]

Due to the nature of no network connectivity, We cannot login (curtin preseed didn't run to setup credentials) or submit logs.

[Scott Moser (smoser)]

heres the diagnosis of the failure here:

a.) the powerVM system was successfully booting and installing xenial, but was failing to bring up networking on the installed system.

b.) curtin had written (as designed) /etc/udev/rules.d/70-persistent-net.rules containing a line like:
  SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="40:f2:e9:5d:be:68", NAME="eth0"

c.) curtin had written an /etc/network/interfaces file like:
   auto eth0
   iface eth0 inet dhcp

d.) the cert team had curtin configuration (late_commands) that was writing /etc/network/interfaces in the target with "dhcp on everything". In order to do this, they read the device names from the install environment via /sys/class/net/ . This resulted in /etc/network/interfaces like:
  auto enP2p1s0f0
  iface enP2p1s0f0 inet dhcp

When the system booted, networking failed to come up because there was no device named 'enP2p1s0f0', as 70-persistent rules named it 'eth0'.

[Jeff Lane  (bladernr)]

Pretty sure this was resolved by GA... should this be marked invalid or some other status to reflect this?