Heat Overlapping ip issue
The hard coded cidr used by heat neutron scenario,
was overlapping with the host systems cidr.
Because the neutron scenario does not used his own router,
it also caused issue for the cfn init scenario.
* The neutron scenario will use the first configured cidr
* The neutron scenario will use his own router
* The neutron scenario will retry singaling the heat-cfn-api,
because if neutron is overload and the vm boot is fast and we
do not wait even for the meta data api service connectivity (enabled
config drive), the signal could be sent in a wrong time when
the L3 connectivity is not ready, but the L2 is.
* Let neutron decide the allocation_pools (simplification)
* Using comma_delimited_list for passing not hard coded dns servers
to the stack
Change-Id: I06bd197b0f6c012a1416016a40f29ddd080b21b9
Partial-Bug: #1297560
4 files changed