While triggering an image at Mentor, it was observed that recipe talloc was breaking at do_configure phase. After a bit of investigation it was figured out that 2 bbclass-es are not taken:
a) qemu.bbclass
b) waf-samba.bbclass

These bbclasses fall under meta-mentor-staging layer ((Reference: http://git.yoctoproject.org/cgit/cgit.cgi/meta-mentor/tree/meta-mentor-staging/classes?h=cedar).

Instead the build was taking waf-samba.bbclass from meta-oe layer and qemu.bbclass from poky/meta layer.

When ran the below in a django shell ( thanks to Michael Wood for the helping hand which helped me understand the problem ):

-------------
(venv)sujith@kdekidd0:~/myproject$ ~/mgc/embedded/mel/2015.12.119/poky/bitbake/lib/toaster/manage.py shell
Python 2.7.6 (default, Jun 22 2015, 17:58:13) 
[GCC 4.8.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
(InteractiveConsole)
>>> from orm.models import *
>>> from bldcontrol.models import BuildRequest
>>> Layer_Version.objects.filter(layer__name="meta-mentor-staging").last().priority
0
>>>  Layer_Version.objects.filter(layer__name="meta-python").last().priority
0
>>>  Layer_Version.objects.filter(layer__name="meta-mentor-staging").first().priority
10
>>> Layer_Version.objects.filter(layer__name="meta-oe").last().priority
0
>>> Layer_Version.objects.filter(layer__name="meta-oe").first().priority
6
>>> 

-------------

I am not sure why the priority is reset to 0 ( zero ). I assume this could be reproducible, if we create multiple layers let's say: meta-a, meta-b, meta-c etc and then copy a bbclass to one of the newly created layers. Update the do_configure as follows
do_configure ()  {
   echo " I am here... " > /tmp/test.log
  #Rest of the code should be kept as it is
  ... 
}

So when user tries to run the recipe's configure from toaster, he/she should be able to see /tmp/test.log file. Else we are hit with a crucial bug.