Vagrant Test Cluster

For development we use Vagrant to set up a Ganeti Cluster to test on.

Setup

The Vagrant Test cluster is not included with Ganeti Web Manager. It is a separate repository that must be cloned:

git clone https://github.com/osuosl/vagrant-ganeti

Once you’ve cloned the repository you can refer to its README for more details on how to use it.

Updating

If you already have the repository, you should also always make sure you are running the most up to date version of the code. There are two ways to do this:

git pull
git submodule update --init

Or you can simply use the update script in the repo which does the previous two commands:

./update

Either way will work, but make sure that both the repository and the submodules are updated.

Basic Usage

The vagrant-ganeti project comes with 3 Ganeti nodes which can be used for testing deployment of VMs. You typically only need 1 node unless you are testing multi-node operations.

  • You can bring up the first node by typing vagrant up node1.

  • The other nodes can be brought up the same way, just swap out node1 with node2 or node3.

You can also bring up every node by using just vagrant up, however if you do not have much RAM or you are already running memory intensive programs, this should be avoided.

Using with Ganeti Web Manager

Once you have node1 running, you can add it to Ganeti Web Manager by using the add cluster button on the cluster page.

The hostname field should be filled in with the IP Address for the given node, which can be found in the Vagrantfile. The username and password are both vagrant, which will let you communicate with the cluster.