Cirros failed to get metadata in boot time with network OVN

[TsutomuNakamura]

In the dashboard, after you select Instances -> [Instance name] -> Log, messages like below are shown.

checking http://169.254.169.254/2009-04-04/instance-id
failed 1/20: up 3.50. request failed
failed 2/20: up 52.55. request failed
failed 3/20: up 101.58. request failed
......

This error still remaining but Cirros is running fine so far.

https://bugs.launchpad.net/openstack-helm/+bug/1745370

https://docs.huihoo.com/openstack/archive/admin-guide-cloud/content/section_metadata-service.html

[TsutomuNakamura]

Is it something related? It might be neccessary in nova.conf.

• /etc/nova/nova.conf

  [neutron]
  ...
  service_metadata_proxy = true
  metadata_proxy_shared_secret = <password>

PY3: nova meta-data 404

[TsutomuNakamura]

Test cases

service_metadata_proxy and metadata_proxy_shared_secret should be declared in /etc/nova/nova.conf in controller or compute node. Check which node should I set.

Set controller	Set compute	Result	Comment
Yes	No	✔️
Yes	Yes	✔️
No	No	✖️
No	Yes	✖️

service_metadata_proxy and metadata_proxy_shared_secret should be set in controller node in version 'mitaka'. Install and configure controller node

service_metadata_proxy and metadata_proxy_shared_secret were NOT set in compute node in version 'mitaka'. Install and configure compute node