This is to hopefully track general issues with devstack bringup. Seems the deployment happens successfully but not without connectivity issues to the appliance VM afterward. For me, I get a VM that gets an address on the mgt network but the VM cannot be reached from the devstack host:
adam@trusty:~$ nova list --all-tenants
+--------------------------------------+-----------------------------------------+----------------------------------+--------+------------+-------------+---------------------------------------------------------------------------------------------+
| ID | Name | Tenant ID | Status | Task State | Power State | Networks |
+--------------------------------------+-----------------------------------------+----------------------------------+--------+------------+-------------+---------------------------------------------------------------------------------------------+
| 71fbb2c7-702d-4fbd-a760-4e06ac2f37f0 | ak-8e435d07-f614-4953-9422-4b8f19556b97 | e9f0651a53004616a2bf6aaf4cb2729e | ACTIVE | - | Running | thenet=fdd6:a1fa:cfa8:6cb2:f816:3eff:fede:ff1a; mgt=fdca:3ba5:a17a:acda:f816:3eff:febb:79c5 |
+--------------------------------------+-----------------------------------------+----------------------------------+--------+------------+-------------+---------------------------------------------------------------------------------------------+
adam@trusty:~$ ip route get fdca:3ba5:a17a:acda:f816:3eff:febb:79c5
fdca:3ba5:a17a:acda:f816:3eff:febb:79c5 from :: dev tap80cebd89-10 src fdca:3ba5:a17a:acda::1 metric 0
cache
I can't ping the VM from the devstack host, however, logging into the VM via console I can ping the link local address on the tap device (fdca:3ba5:a17a:acda::1). It doesn't appear that anything is being dropped by iptables on the host or the VM.
I've confirmed the packet is getting dropped in the fallback chain. The reverse test from VM to host works because the related rule allows the traffic through.