62e02fdbf4
Due to a revert in the Docker code, thin neutron containers do not work
at this time. The neutron plugin agents do not need to exist in the same
container as the other agents however. This allows for using either
linuxbridge or openvswitch
The docker commit that did the revert:
|
||
|---|---|---|
| .. | ||
| cinder-api-scheduler.yml | ||
| cinder-backup.yml | ||
| cinder-volume.yml | ||
| designate.yml | ||
| galeradb.yml | ||
| glance-api-registry.yml | ||
| heat-api-engine.yml | ||
| horizon.yml | ||
| keepalived.yml | ||
| keystone.yml | ||
| magnum-api-conductor.yml | ||
| mariadb.yml | ||
| neutron-agents.yml | ||
| neutron-linuxbridge-agent.yml | ||
| neutron-server.yml | ||
| nova-api-compute-network.yml | ||
| nova-api-conductor-scheduler-consoleauth-novncproxy.yml | ||
| nova-compute-network.yml | ||
| nova-compute.yml | ||
| rabbitmq.yml | ||
| README.md | ||
Docker compose
These scripts and docker compose files can be used to stand up a simple installation of openstack. Running the 'tools/genenv' script creates an 'openstack.env' suitable for running on a single host system as well as an 'openrc' to allow access to the installation.
Once you have run that you can either manually start the containers using the 'docker-compose' command or try the 'tools/kolla start' script which tries to start them all in a reasonable order, waiting at key points for services to become available. Once stood up you can issue the typical openstack commands to use the installation. If using nova networking use:
# source openrc
# tools/init-runonce
# nova boot --flavor m1.medium --key_name mykey --image puffy_clouds instance_name
# ssh cirros@<ip>
Else if using neutron networking use:
# source openrc
# tools/init-runonce
# nova boot --flavor m1.medium --key_name mykey --image puffy_clouds instance_name --nic net-id:<net id>
# ssh cirros@<ip>