OpenStack (Rocky): VM not spawning on provider network though it is successfully getting spawned on self-service network.
Hello Everyone, I am trying to set-up OpenStack (Rocky) cloud using the standard guide (Available at link:https://docs.openstack.org/install-guide/) for OpenStack (Rocky) installation. Below here are the details provided. Please let me know if any further details are required, I shall be very happy to provide.
Cloud Setup
My setup contains one controller node and two compute nodes. Both the nodes contain 2 NIC. One for management network (10.0.0.0/24 - NIC name enp4s0) and other for provider network (10.20.4.0/22 - enp2s0).
Controller node: All the services are functional and running successfully on the controller node. systemctl status SERVICE
shows all the services required for successful working of controller node are up and running. Neutron (networking service) is also running on the controller node.
Compute nodes: All the services are functional and running successfully on the compute node. systemctl status SERVICE
shows all the services required for successful working of compute node up and running.
I have two networks created. 1. Provider network (flat with external provider network - 10.20.4.0/22) 2. Self-service network (192.168.0.0/24)
There already is 1 OpenStack cloud running on the same physical network (The physical network I am using for the provider network) if that can be a problem which I think is very unlikely. Both the setups are using separate IP pools.
Problem
Instances are successfully getting created on the Self-service network. But when I create an instance on Provider network, instance gets with an error "Failed to allocate the network(s), not rescheduling". After trying to solve the problem, I think (Just a guess though :-) ) I have figured out the problem.
The problem is with the bridge that is getting created on the physical provider network interface (brqeedc5776-03 on enp2s0) on both compute nodes. The same bridge that that is getting created on the physical provider network interface on controller is working fine. Physical provider network interface (enp2s0) gets added to the bridge automatically on controller node and is working fine whereas on the compute node it is missing as can be seen in the following screenshots.
image:/home/abdul/brctl-show-controller.xcf Physical interface 'enp2s0' already added to the bridge-brqeedc5776-03 (This bridge is being used for provider network).
image:/home/abdul/brctl-show-compute01.xcf VXLAN interface successfully added to the bridge and 1 instance successfully running on that bridge-brqab88d4ca-68 (This bridge is being used for Self-service network).
Solution that did not help
I tried adding the physical interface to the bridge-brqeedc5776-03 (This bridge is being used for provider network) manually using brctl addif brqeedc5776-03 enp2s0
on compute node. But it did not solve the problem and I am still getting the same error.
Log files: I am not getting any error in log files on controller node. The only error I am getting is in the logfile on path /var/log/neutron/neutron-linuxbridge-agent.log on the controller node. I get this error only on the controller node on which the VM ...