Launch VM fails on Compute when VXLAN ML2 type driver is used [closed]

asked 2014-04-07 02:37:47 -0500

KP gravatar image

I have two Openstack nodes (controller & Compute and a Compute). VM's are getting launched fine on the node that also acts as the controller. But, the VM's that are scheduled on the compute node seems to go to error state and the error I see on Nova screen is attached at the end. I am running Icehouse version and my ML2 type driver is VXLAN. I used devstack for my installation. If I use non-tunnel mode and change it to VLAN, I don't see this error and VM's are launched fine in compute nodes as well.

My nova.conf in compute does have the following: neutron_url = http://192.28.10.149:9696 neutron_admin_auth_url = http://192.28.10.149:35357/v2.0

Ofcourse, the above IP addresses of controller are reachable from compute.

This is what I have in my ml2_conf.ini in my controller. [ml2] tenant_network_types = vxlan type_drivers = vxlan mechanism_drivers = openvswitch

I assume this file on compute is not needed.

2014-04-06 10:49:45.906 TRACE nova.compute.manager [instance: efede8e5-6bdb-4c2d -8e77-24a45dc7f652] File "/opt/stack/python-neutronclient/neutronclient/v2_0/c lient.py", line 1315, in _pagination 2014-04-06 10:49:45.906 TRACE nova.compute.manager [instance: efede8e5-6bdb-4c2d-8e77-24a45dc7f652] res = self.get(path, params=params) 2014-04-06 10:49:45.906 TRACE nova.compute.manager [instance: efede8e5-6bdb-4c2d-8e77-24a45dc7f652] File "/opt/stack/python-neutronclient/neutronclient/v2_0/client.py", line 1288, in get 2014-04-06 10:49:45.906 TRACE nova.compute.manager [instance: efede8e5-6bdb-4c2d-8e77-24a45dc7f652] headers=headers, params=params) 2014-04-06 10:49:45.906 TRACE nova.compute.manager [instance: efede8e5-6bdb-4c2d-8e77-24a45dc7f652] File "/opt/stack/python-neutronclient/neutronclient/v2_0/client.py", line 1280, in retry_request 2014-04-06 10:49:45.906 TRACE nova.compute.manager [instance: efede8e5-6bdb-4c2d-8e77-24a45dc7f652] raise exceptions.ConnectionFailed(reason=_("Maximum attempts reached")) 2014-04-06 10:49:45.906 TRACE nova.compute.manager [instance: efede8e5-6bdb-4c2d-8e77-24a45dc7f652] ConnectionFailed: Connection to neutron failed: Maximum attempts reached 2014-04-06 10:49:45.906 TRACE nova.compute.manager [instance: efede8e5-6bdb-4c2d-8e77-24a45dc7f652]

Pls let me know if you need more information.

My localrc on Controller is (relevant ones that i thought):

enable_service q-l3

enable_service quantum

Q_PLUGIN=ml2

Q_ML2_PLUGIN_MECHANISM_DRIVERS=openvswitch

Q_ML2_TENANT_NETWORK_TYPE=vxlan

Q_ML2_PLUGIN_TYPE_DRIVERS=vxlan

Q_PLUGIN_EXTRA_CONF_PATH=/home/user1/ML2/devstack

ENABLE_TENANT_VLANS=False

PHYSICAL_NETWORK=physnet2

OVS_PHYSICAL_BRIDGE=br-eth3

ENABLE_TENANT_TUNNELS=True

localrc on compute is:

Q_PLUGIN=ml2

Q_ML2_TENANT_NETWORK_TYPE=vxlan

ENABLE_TENANT_TUNNELS=True

Thanks, Paddu

edit retag flag offensive reopen merge delete

Closed for the following reason question is not relevant or outdated by rbowen
close date 2016-09-21 15:57:07.267635

Comments

Closing: This references a version of OpenStack which has been marked End Of Life - https://releases.openstack.org/ Please open a new question if you're still experiencing this with a more recent version of OpenStack.

rbowen gravatar imagerbowen ( 2016-09-21 15:57:01 -0500 )edit