One time setup for use in Grid’5000#

Overview#

In this notebook we’ll install EnOSlib in a dedicated virtual environment and use it in the subsequent notebook.

Before we proceed, please keep in min the following representation - Every single cell of the various notebooks will be applied on the frontend node of the chosen site (or a dedicated node if it applies) - Each site frontend host a dedicated scheduler instance (so you can reserve, delete, view you resources from this node) - By default you can’t reach nodes inside the Grid’5000 network from the internet (Various options exist to overcome this but are out-of-the scope of this notebook)

g5k_topology

Installing the latest version of EnOSlib#

This installs EnOSlib (+ some extra libs) in a virtual environment and install a new Jupyter kernel to be used in all the notebooks

[ ]:

%%bash

virtualenv -p python3 my_venv
source my_venv/bin/activate

# install a specific major version
python3 -m pip install -U 'enoslib[jupyter]>=10,<11'

# install the latest development version
# python3 -m pip install -U 'enoslib[jupyter] @ git+https://gitlab.inria.fr/discovery/enoslib'

# install some extra dependencies
python3 -m pip install matplotlib seaborn pandas scapy

# make sure we can use the python interpreter of the venv as a jupyter kernel
python3 -m pip install ipykernel
python3 -m ipykernel install --user --name my_venv

Activate this environment#

Goal: Make sure we’re using the library installed in the previous virtual environment.

  • Switch the kernel to my_venv: in the notebook interface, go to Kernel > Change kernel and select my_venv

  • Refresh the browser (optional but sometime required)

  • Re-execute this notebook

Checking the library#

The following must succeed after switching to the new kernel my_venv

[ ]:

import enoslib as en

# check the version
en.check()

Special case if you are running this notebook on a dedicated node#

Regarding the access to the REST API, a node on G5k is considered to belong to the external world. So you need to set your credentials to be able to connect to the REST API.

[ ]:

from grid5000.cli import auth

# CHANGE THE LOGIN!
auth("msimonin")