When today i have added new compute node and noticed following error on neutron-server logs
May 20 13:38:47 os-infra-1-neutron-server-container-49dba0f4 neutron-server[87]: 2021-05-20 13:38:47.668 87 ERROR neutron.plugins.ml2.managers [req-38f48ec1-7395-4a54-8884-39c99e6e09ee ff1b8371df904ab1a01a46294d985303 4852c36c80354e30b177b02fb99b567a - default default] Failed to bind port 0ad9de40-22ee-4ddc-b6b1-c99902e8a383 on host os-compute-1.example.net for vnic_type normal using segments [{'id': 'd02daa50-7c36-4fdd-989b-47e24f762943', 'network_type': 'geneve', 'physical_network': None, 'segmentation_id': 1, 'network_id': 'b96364e8-340b-4024-97ff-2796be2acab6'}]
root@os-compute-1:~# hostname
os-compute-1.example.net
# ovs-vsctl list open_vswitch | grep external_ids
external_ids : {hostname=os-compute-1, ovn-bridge-mappings="vlan:br-provider", ovn-chassis-mac-mappings="vlan:aa:bb:cc:dd:ee:01", ovn-cms-options=enable-chassis-as-gw, ovn-encap-ip="192.168.40.3", ovn-encap-type="geneve,vxlan", ovn-remote="tcp:172.30.40.2:6642", rundir="/var/run/openvswitch", system-id="57c973ee-464a-40df-8436-1d046a69a671"}
## Solution
root@os-compute-1:~# ovs-vsctl set open_vswitch . external_ids:hostname="os-compute-1.example.net"
I can see we have following patch to fix this issue - https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/620604/1/tasks/providers/setup_ovs_ovn.yml#18
but my question is why are we using short name instead use FQDN from real hostname command from OS?
This bug is a duplicate of https:/ /bugs.launchpad .net/openstack- ansible/ +bug/1922978. Please see URL for patch.