Activity log for bug #1971488

Date Who What changed Old value New value Message
2022-05-03 20:10:11 Juan Pablo Suazo bug added bug
2022-05-03 20:53:23 OpenStack Infra kolla-ansible: status New In Progress
2022-05-05 21:42:34 Juan Pablo Suazo kolla-ansible: status In Progress New
2022-05-06 13:11:06 OpenStack Infra kolla-ansible: status New In Progress
2022-05-06 13:12:31 Radosław Piliszek kolla-ansible: importance Undecided Low
2022-05-06 13:12:56 Radosław Piliszek summary OpenVSwitch playbook does not allow controllers without OVS OpenVSwitch playbook does not allow enough flexibility in host roles assignment
2022-05-06 13:13:19 Radosław Piliszek description Tasks `Set system-id` an `Ensuring OVS bridge is properly setup` from the `ansible/roles/openvswitch/tasks/post-config.yml` assumes controller nodes are also network nodes. If this is not the case, the tasks attempt to run commands against OpenVSwitch containers that are not present, stopping the deployment of openstack. - name: Set system-id become: true command: docker exec openvswitch_vswitchd ovs-vsctl set Open_vSwitch . external_ids:system-id={{ openvswitch_system_id }} - name: Ensuring OVS bridge is properly setup become: true command: docker exec openvswitch_db /usr/local/bin/kolla_ensure_openvswitch_configured {{ item.0 }} {{ item.1 }} register: status changed_when: status.stdout.find('changed') != -1 when: - inventory_hostname in groups["network"] or (inventory_hostname in groups["compute"] and computes_need_external_bridge | bool ) with_together: - "{{ neutron_bridge_name.split(',') }}" - "{{ neutron_external_interface.split(',') }}" Found this bug in our victoria deployment when configuring some nodes to be ONLY control nodes. As this nodes do not have OpenVSwitch services running on them, this tasks fail when attempting to run this commands against their containers. Tasks `Set system-id` an `Ensuring OVS bridge is properly setup` from the `ansible/roles/openvswitch/tasks/post-config.yml` assumes controller nodes are also network nodes. If this is not the case, the tasks attempt to run commands against OpenVSwitch containers that are not present, stopping the deployment of openstack. - name: Set system-id   become: true   command: docker exec openvswitch_vswitchd ovs-vsctl set Open_vSwitch . external_ids:system-id={{ openvswitch_system_id }} - name: Ensuring OVS bridge is properly setup   become: true   command: docker exec openvswitch_db /usr/local/bin/kolla_ensure_openvswitch_configured {{ item.0 }} {{ item.1 }}   register: status   changed_when: status.stdout.find('changed') != -1   when:     - inventory_hostname in groups["network"]       or (inventory_hostname in groups["compute"] and computes_need_external_bridge | bool )   with_together:     - "{{ neutron_bridge_name.split(',') }}"     - "{{ neutron_external_interface.split(',') }}" Found this bug in our victoria deployment when configuring some nodes to be control+loadbalancer(network) nodes. As this nodes do not have OpenVSwitch services running on them, this tasks fail when attempting to run this commands against their containers.