Deployment scripts for the SMO project

These scripts are designed to deploy the NEPHELE infrastructure, the SMO (Service Management Operations) and any additional demo on a server. They PyInfra for deployment and assumes you have a working SSH connection to the target server.

Available scripts:

• deploy-docker.py
• eploy-brussels.py
• deploy-mk8s.py
• deploy-infra.py

Script deploy-docker.py

Deploy Docker and Docker registry on a target host, from official images of docker.com.

Command:

pyinfra -y -vvv --user USER HOST deploy-docker.py

• Target host (HOST) is expected to bee Ubuntu or Debian.

• A pre existing account (USER) with ssh acces is required.

What the script does:

• Check that the server is Ubuntu or Debian

• Install som base packages (curl, gpg, ...)

• Install docker.com key and source list of packages

• Install packages: docker-ce, docker-ce-cli, containerd.io

• add USER to the docker group (allowing use of docker commands without need of sudo)

• Install and start the official registry container. The registry is in insecure mode (no access restriction) on port 5000 by default (configure this port with REGISTRY_PORT constant if needed).

Caveats:

• if some previous installation of Ubuntu package for docker, it will be necessary to remove them (and their stored images). See commented lines around line 93: "dpkg --purge --force-all containerd docker.io runc"

Script deploy-brussels.py

Deploy the SMO code base and build the brussels docker images. The images are then pushed on the local docker registry. You need to ensure an existing docker environment (ie. use deploy-docker.py).

Command:

pyinfra -y -vvv --user USER HOST deploy-brussels.py

• Target host (HOST) is expected to bee Ubuntu or Debian.

• A pre existing account (USER) with ssh acces is required.

What the script does:

• Check that the server is Ubuntu or Debian

• Install some base packages (curl, gpg, ...)

• Install uv for the USER

• Download and install with uv the SMO code source. By default the Nephele SMO version, see SMO and SMO_URL constants to use another source repository.

• Download and install the hdarctl binary (this implies the use of a amd64 host).

• Build and install the 3 brussels demo images, push them on the local docker registry. The registry port (5000 by default) can be changed, see REGISTRY_PORT constant.

Caveats:

• The docker registry is accessed via LOCAL_IP constant, set by default to 127.0.0.1. This value can be changed to any address.

Script deploy-mk8s.py

Deploy a minimal microk8s environment on a Ubuntu like distribution and activate helm and prometheus on it. microk8s is installed through the snap package system that contains preconfigured services (including helm and prometheus).

Command:

pyinfra -y -vvv --user USER HOST deploy-mk8s.py

• Target host (HOST) is expected to bee Ubuntu or Debian.

• A pre existing account (USER) with ssh acces is required.

What the script does:

• Check that the server is Ubuntu or Debian

• Install some base packages (curl, gpg, ...)

• Install the snapd package

• Install the microk8s package, enable prometheus plugin (helm is already activated)

• Dump the microk8s configuration in local file ~/.kube/config. This config file can be used to configure later a karmada server.

Caveats:

• Prometheus is annouced deprecated, however an equivalent tool is proposed.

Script deploy-infra.py

Not finished yet. This is a work in progress and not ready for production use.

Deploy the full Nephele project tools. This script tries to follow the Nephele project guidelines.

It's main goal is to test an installation of the karmada meta ochestrator.

Background information on how to install the platform (manually): https://gitlab.eclipse.org/eclipse-research-labs/nephele-project/nephele-platform

See also https://pad.cloud.abilian.com/-JKefCz7S3abBQhSbv1AIg# for some notes on the available installation documentation from NEPHELE.

How to Run

1. Install

Run uv sync then activate the virtual environment (typically source .venv/bin/activate).

2. Setup (optional)

Update inventory.py with the correct SSH user and host information. The script assumes you have SSH access to the target server(s).

You may also call pyinfra directly with the ip address or the hostname of the server you want to deploy to, as described below.

3. Run the Deployment Script

• If targeting a single host by IP directly (e.g., 157.180.84.240) with user admin:

  pyinfra 157.180.84.240 -y -v --user root deploy-infra.py

  In this case, host.name will be 157.180.84.240, and TARGET_IP_FOR_MAKEFILE will use this.

• If using inventory.py (e.g., the smo_server_alias example from inventory.py comments): First, ensure inventory.py has something like:

  # inventory.py
  hosts = {
      'smo_server_alias': {
          'ssh_host': 'actual_connect_ip_or_dns',
          'ssh_user': 'your_admin_user_with_sudo',
          'public_ip': '157.180.84.240' # The IP for the Makefile
      }
  }

  Then run:

  pyinfra inventory.py smo_server_alias deploy.py

  Or to target all hosts in the inventory:

  pyinfra inventory.py deploy.py

  In this case, host.data.get('public_ip', host.name) will pick up '157.180.84.240'.

• Debugging: add "-vv", "-vvv" and/or "--debug" to the command line to get more verbose output. This is useful for debugging and understanding what the script is doing. See https://docs.pyinfra.com/en/3.x/cli.html#additional-debug-info for more details.