neutron-openvswitch-agent crashing: Bridge br-ex for physical network physnet1 does not exist. Agent terminated!

[Markus Adam]

Bug description:
Deployed Kolla Ansible all-in-one Pike on CentOS.
enable_neutron_dvr and enable_neutron_provider_networks enabled.
Observed that container neutron-openvswitch-agent keeps restarting.
Seems that Neutron expects br-ex bridge which is not present and not configured neither by Kolla Ansible neither manually in my case.
Documentation is sparse when it comes to provider networks:
https://docs.openstack.org/kolla-ansible/latest/reference/networking-guide.html
I'm not really sure whether br-ex should be configured by Kolla or manually?

Bug details:
# tail /var/lib/docker/volumes/kolla_logs/_data/neutron/neutron-openvswitch-agent.log
2018-02-19 08:12:15.964 7 INFO neutron.plugins.ml2.drivers.openvswitch.agent.ovs_neutron_agent [req-4683d9e3-c252-4d4c-934f-af1cc7f88366 - - - - -] Mapping physical network physnet1 to bridge br-ex
2018-02-19 08:12:15.965 7 ERROR neutron.plugins.ml2.drivers.openvswitch.agent.ovs_neutron_agent [req-4683d9e3-c252-4d4c-934f-af1cc7f88366 - - - - -] Bridge br-ex for physical network physnet1 does not exist. Agent terminated!

# ip link
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP mode DEFAULT qlen 1000
    link/ether 52:54:00:6a:92:9c brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP mode DEFAULT qlen 1000
    link/ether 52:54:00:4f:13:a9 brd ff:ff:ff:ff:ff:ff
4: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN mode DEFAULT
    link/ether 02:42:1e:f1:b3:a2 brd ff:ff:ff:ff:ff:ff
5: ovs-system: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether d6:45:54:f3:e5:4e brd ff:ff:ff:ff:ff:ff
6: br-int: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether a2:cb:56:5f:28:4f brd ff:ff:ff:ff:ff:ff

[Markus Adam (markus.adam)]

Setting enable_neutron_dvr: "no" did the job.
It appears that enable_neutron_dvr: "no" is not working in all-in-one deployments.

[Markus Adam (markus.adam)]

It appears that enable_neutron_dvr: "yes" is not working in all-in-one deployments.

[Markus Adam (markus.adam)]

Having set:
enable_neutron_provider_networks: "yes"
enable_neutron_dvr: "no"

Makes br-ex configured in all-in-one

# ip link
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP mode DEFAULT qlen 1000
    link/ether 52:54:00:6a:92:9c brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master ovs-system state UP mode DEFAULT qlen 1000
    link/ether 52:54:00:4f:13:a9 brd ff:ff:ff:ff:ff:ff
4: docker0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN mode DEFAULT
    link/ether 02:42:1e:f1:b3:a2 brd ff:ff:ff:ff:ff:ff
5: ovs-system: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether d6:45:54:f3:e5:4e brd ff:ff:ff:ff:ff:ff
11: br-ex: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether 0a:2e:8d:bc:2a:42 brd ff:ff:ff:ff:ff:ff
12: br-int: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether 9e:35:a2:99:39:44 brd ff:ff:ff:ff:ff:ff
13: br-tun: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether ea:63:29:1e:02:40 brd ff:ff:ff:ff:ff:ff

[Robert (robert4)]

all-in-one deployment ,
This in globals.yml does not fix the problem:
enable_neutron_provider_networks: "yes"
enable_neutron_dvr: "no"

[Michal Nasiadka (mnasiadka)]

Pike is long EOL