[Bug 1639202] Re: Cannot enlist/commission machines in MAAS 2.1 with usb network adapter

[Dave Chiluk]

Alright so the problem at present appears to be that the machine is pxe
booting off of a nic with a mac address that is not showing up after the
kernel boots.

The way the boot works is the bios/efi launches a pxe network stack.
This typically makes a dhcp request.  The DHCP server responds with an
IP address, and the address of the PXE/TFTP server *(in this case the
maas server).  The network stack firmware on the client then requests
the kernel, initramfs and kernel arguments from the PXE server.  The
bios/efi pxe network stack then downloads this, and executes the kernel.

One of the arguments maas is responding with BOOTIF=01-9c-
eb-e8-3c-52-cc.  This means the original pxe request originates from
this mac address.  When the initramfs starts it runs a script function
called configure_networking that attempts to set up the BOOTIF=01-9c-
eb-e8-3c-52-cc NIC, but it doesn't appear to exist to the OS.

This could mean a few things.  
- The NIC doing the initial pxe request is different than the usb-c one.  Is there a chance that there's a wireless nic that has a pxe stack that you've configured?  I know some newer machines are able to pxe boot off of their network cards so this would be useful to check.
- The mac address is changing between the pxe request and the OS boot.
- IPv6 is in the mix.  Are you attempting to boot via ipv6?
- The PXE server is responding with the incorrect mac address in BOOTIF.
The last two can be checked by looking at /var/log/rackd.log on your maas server.  You should be able to grep for 01-9c-eb-e8-3c-52-cc or 01-84-7b-eb-55-c1-95 in the rackd.log to see which nic is making the pxe request.  If 01-9c-eb-e8-3c-52-cc shows up in the rackd.log then it's pretty definitive that the issue is booting using a nic with that mac somehow. 

Please check the above and let me know what you discover.

Thanks,
Dave.

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to initramfs-tools in Ubuntu.
https://bugs.launchpad.net/bugs/1639202

Title:
  Cannot enlist/commission machines in MAAS 2.1 with usb network adapter

Status in MAAS:
  Invalid
Status in maas-images:
  New
Status in initramfs-tools package in Ubuntu:
  New
Status in linux package in Ubuntu:
  Invalid

Bug description:
  After upgrade from MAAS 2.0 to MAAS 2.1:

  Cannot enlist/commission client machines via network boot (or pre-
  staging machine) when using usb-c network adapter D59GG (e.g.
  Precision 5510). This was working ok with MAAS 2.0.

  Using the Xenial 16.04 base image for enlist/commission (no minimum
  kernel set).

  Enlist/Commission/Deploy works fine with other laptop and desktop
  models which have a built in NIC.

  From an already deployed Precision 5510 system (16.04.1) I can see the
  following module is loaded when the usb-c adapter is connected (and
  working):

  $ lsmod
  usbnet	45056 1 cdc_ether

  
  Errors received on client during enlisting with MAAS:

  no /run/net-bootif.conf
  lvmetad is not activated
  ...invalid path for logical volume.
  gave up waiting for root device
  common problems: boot args (cat /proc/cmdline)
  check rootdelay = (did system wait long enough)
  check root = (did the system wait for the right device?)
  missing modules (cat /proc/modules)
  ALERT!  /dev/disk/by-path/ip-<ipaddress>:3260-iscsi-iqn,2004.05.com.ubuntu:maas:ephemeral-ubuntu-amd64-generic-xenial-daily-lun-1 does not exist. Dropping to shell!
  usbcore: registered new interface driver usbhid
  usbhid: USB HID core driver

  - cat /proc/cmdline shows mac address that of usb-c network adapter.
  - cat /proc/modules includes “usbnet 45056 cdc_ether, Live 0xffffffffc009c000”
  - cat /proc/modules | grep usb … includes “ usbnet 45056 and usbhid 49152”

  
  I've tried enlisting with older boot image and different kernel versions (14.04 and 16.04 with ga-16.04, hwe-16.04 set) but get the same problem.

  It seems like the usb-c network adapter isn't loading properly or
  maybe just not quickly enough?

  Please let me know if you require any more info. I can provide info
  from /var/log/maas/* and dpkg -l '*maas*'|cat if need be.

To manage notifications about this bug go to:
https://bugs.launchpad.net/maas/+bug/1639202/+subscriptions