This script deploys an OpenStack on Grid'5000 using Kolla, isolating some services on different nodes. Everything must be configured in a single JSON file and then the deployment can be run in one command.
Kolla-G5K-Multinode requires pip, Execo and Jinja.
To install Kolla-G5K-Multinode, first clone the repository:
https://github.com/BeyondTheClouds/kolla-g5k
Then install the dependencies:
cd kolla_g5k
pip install -r requirements.txt --user
Make sure to add ansible binaries to your PATH (export PATH=YOUR_HOME/.local/bin:$PATH).
To deploy a full stack, you need to provide a configuration file in the YAML format. This
file must include information about the Grid'5000 job and how the OpenStack services must
be deployed. reservation.yaml.sample is given as an example.
# copy the sample file
cp reservation.yaml.sample reservation.yaml
#edit it with your parameters
<editor> reservation.yaml
Then, to run launch the deployment, run:
./01_deploy_nodes.py
by default the script use the file
reservation.yamlas configuration file
On the first run, the script will create an OAR Grid job with a specific name. On subsequent runs, it will find the running job and use it. The job name can be specified in the YAML configuration file.
Run ./01_deploy_nodes.py -h for a full list of command-line arguments.
After the execution, the ./current directory will contain some generated
files (please note that current is a symbolic link toward the real directory
associated to your deployment)
- The network interfaces inserted in the
templates/globals.yml.jinja2file are taken from the Grid'5000 API for the first cluster specified in the YAML file. This may be a problem if you reservation contains multiple Grid'5000 clusters composed of nodes that have different network device names, e.g. paravance that haseth0andleth1, and parapluie that has deviceseth1andeth2.
The deployment makes use of a private docker registry configured as a mirror of the official docker registry. There are two modes :
ceph: false. It will start a fresh registry that will cache the images for the duration of the experimentceph: true. It will use an existing Ceph rados block device of the Rennes cluster. The block device will be mounted and used as the registry storage. Setting this is usefull as the cache will persist different experiments.
The G5k ceph tutorial will guid you on how to create your rados block device.
This code is released under the GNU General Public License.