Development

Machines deployment

In the Quick start section we have seen how to deploy a manager virtual machine. The development system is currently made of only the manager virtual machine. In the future some machines could be added to the deployment procedure. To deploy all of them, it is sufficient to pass to the discos-deploy script the discos cluster instead of a single machine name, the discos keyword always identifies the whole environment, so it will install all the machines found in the corresponding Ansible inventory.

$ discos-deploy discos:development

This command will connect via SSH to all development machines and provision the whole system (create users, configure networking, install yum packages, configure NIS, Lustre, ACS and its dependencies, install some utilities, and finally the DISCOS dependencies).

Working without Vagrant

If you wish to deploy a development environment on physical machines, or you simply want to install the virtual machines by yourself using VirtualBox or another virtual machine provider, you can still use the discos-deploy script to perform the automatic deployment procedure. You first have to setup the machines’ network interfaces, in order to enable Ansible access to them via the correct IP address, and then start the deployment procedure by passing the --no-vagrant, -n flag to the discos-deploy script:

$ discos-deploy manager:development --no-vagrant

This will let the script know that you do not want to create any (other) virtual machine on your system, and that you want to apply the whole configuration onto an existing, physical or virtual, set of machines.

DISCOS setup

Manual setup

To download the DISCOS control software, you can use the discos-get command, covered in the Get a DISCOS branch section.

Once the download of the repository is completed, you have to perform the setup procedure like we have seen in the Install DISCOS section.

Automatic setup

You can let the deployment procedure perform the DISCOS control system setup for you by specifying the --branch argument to the discos-deploy script, followed by the branch name you want to deploy and install. Since the procedure will internally call the discos-get script for you, if you are deploying in a development environment, you also have to provide to the script the --station argument, followed by the name of the desired station.

$ discos-deploy discos:development --tag discos1.0-rc02 --station noto

You can choose a station among medicina, noto and srt. The discos-get command gets executed on all machines of the system at the same time.