Deploying Mirantis OpenStack 6.1 results in network configuration issues with br-aux

Bruce, please add fuel version and diagnostic snapshot?

There has been a zillion reinstalls after that last snapshot. I’ll upload another snapshot, to that same ticket. I’m on the 10th or so iteration of “OS201p"
Now all the nodes are running Ubuntu.

**** Neutron with VLAN segmentation

** Public
IP Range 10.34.88.21 to 10.34.89.220
CIDR 10.34.88.0/22
Use VLAN tagging Yes, 1002
Gateway 10.34.88.1
Floating IP Range 10.34.90.21 to 10.34.91.220

** Storage
CIDR 192.168.1.0/24
Use VLAN tagging Yes, 102

** Management
CIDR 192.168.0.0/24
Use VLAN tagging Yes, 101

** Neutron L2 Configuration
VLAN ID Range 1003 to 1004
Base MAC Address fa:16:3e:00:00:00

** Neutron L3 Configuration
Internal Network CIDR 192.168.111.0/24
Internal Network Gateway 192.168.111.1
Guest OS DNS Servers 10.34.64.11 & 10.34.64.12

Please look through new Fuel snapshot once uploaded by Mark Miller.

I split the upload to 3 parts
fuel, controller90, controller92, and the computenodes 84,85,86,87,88,89

The problem is to get a 2nd external vlan 205 10.34.72.0/22 to access DBs on Layer2
+--------------------------------------+-----------+-------------------------------------------------------+
| id | name | subnets |
+--------------------------------------+-----------+-------------------------------------------------------+
| 9a66fc1b-0e12-4f70-9c6f-87b50e13b698 | net04_ext | 7061a3ea-ebd5-414d-8137-c72f6d23ec97 10.34.88.0/22 |
| 00bd4dce-7945-459f-bd83-94b42fb2a7d4 | net04 | 178e3239-8595-4062-9986-72b5df2d5c8c 192.168.111.0/24 |
+--------------------------------------+-----------+-------------------------------------------------------+

timed out
put 3 files here:
https://drive.google.com/folderview?id=0ByPFLm_01E_EZXd5cUc1a29yY1E&usp=sharing

if you need individual links for wgets...
https://drive.google.com/file/d/0ByPFLm_01E_EeG80eXh1SmdvYTg/view?usp=sharing
https://drive.google.com/file/d/0ByPFLm_01E_EQ2d2bzlIbVJRM0U/view?usp=sharing
https://drive.google.com/file/d/0ByPFLm_01E_EU3gyTFhOdmRubjQ/view?usp=sharing

Mark, could you please clarify some moments in your setup.

dmidecode.txt files for nodes 90 and 92 show that both of nodes are hardware Dell servers.
In your setup only Fuel Master node operates inside VMware virtual machine, so you
do not use vCenter as hypevisor, only KVM.

During deployment there are no errors in puppet logs.

I draw a diagram of your network setup http://paste.openstack.org/show/475904/

I do not understand about your question access to DB on Layer2.

In your setup each tenant traffic will be tagged (according to settings it will be VLANS 1003 or 1004).
When VM emits traffic compute nodes which hosts this VM will mark this with appropriate
VLAN tag and send it into br-prv, which then will be passed through br-aux and bond0.

When controller receives traffic from compute node it will accept it will pass it through
bond0 - br-aux - br-prv. Neutron L3 agent on controller have to route this traffic out
via br-ex (in you case it is VLAN 1002). For you setup VLAN 1002 is considered as provider
network (this is network 'net04_ext' from tenant perspective).

If you are trying to reach external network with VLAN 205 from VLAN 1002 I suppose that
there is not L3 connectivity in your network infrastructure.

Please clarify what you are trying to achieve.

Attaching network scheme, so it will be easy to reference to it
network_scheme:
   transformations:
   - action: add-br
     name: br-fw-admin
   - action: add-br
     name: br-mgmt
   - action: add-br
     name: br-storage
   - action: add-br
     name: br-ex
   - action: add-br
     name: br-floating
     provider: ovs
   - action: add-patch
     bridges:
     - br-floating
     - br-ex
     provider: ovs
     mtu: 65000
   - action: add-br
     name: br-prv
     provider: ovs
   - action: add-br
     name: br-aux
   - action: add-patch
     bridges:
     - br-prv
     - br-aux
     provider: ovs
     mtu: 65000
   - action: add-port
     bridge: br-fw-admin
     name: eth0
   - bridge: br-aux
     name: bond0
     interfaces:
     - eth1
     - eth2
     - eth4
     - eth5
     - eth6
     bond_properties:
       lacp_rate: fast
       mode: 802.3ad
       xmit_hash_policy: layer2+3
     interface_properties:
       vendor_specific:
         disable_offloading: true
     action: add-bond
   - action: add-port
     bridge: br-storage
     name: bond0.102
   - action: add-port
     bridge: br-mgmt
     name: bond0.101
   - action: add-port
     bridge: br-ex
     name: bond0.1002

No updates for 'Incomplete' bug. Moving to 'Invalid'. If you provide more info, please reopen the bug.