add support for CernVM-FS, so software repositories like EESSI can be mounted easily

[boegel]

If CernVM-FS would be available on the nodes, then software repositories like EESSI could be mounted easily, which would make a CitC cluster even more like a real cluster, since it would basically come with (scientific) software (properly optimized) out like GROMACS, R, TensorFlow, OpenFOAM, etc. of the box... 😲 .

I think this can mostly be done via the compute_image_extra.sh script to provision nodes, so it may be just a documentation problem more than anything else (or perhaps provide a separate script that can be called from compute_image_extra.sh to opt-in to it).

I'll try and get this working myself, and post more info in this issue, so you can make an informed decision about the best way forward for this...

[boegel]

By adding these commands to compute_image_extra.sh, and then running sudo /usr/local/bin/run-packer, the workernodes have access to the EESSI pilot repository:

# install CernVM-FS
sudo dnf install -y  https://ecsft.cern.ch/dist/cvmfs/cvmfs-release/cvmfs-release-latest.noarch.rpm
sudo dnf install -y cvmfs
sudo dnf install -y https://github.com/EESSI/filesystem-layer/releases/download/v0.2.3/cvmfs-config-eessi-0.2.3-1.noarch.rpm

# configure CernVM-FS (no proxy, 10GB of quota for CernVM-FS cache)
sudo bash -c "echo 'CVMFS_HTTP_PROXY=DIRECT' > /etc/cvmfs/default.local"
sudo bash -c "echo 'CVMFS_QUOTA_LIMIT=10000' >> /etc/cvmfs/default.local"
sudo cvmfs_config setup

To check that it's working (after rebuilding the node image with run-packer):

• start interactive job:

  srun --pty /bin/bash

• source EESSI init script to set up environment:

  source /cvmfs/pilot.eessi-hpc.org/latest/init/bash

• check available modules:

  module avail openfoam gromacs tensorflow

• load a module and go!

  module load GROMACS/2020.1-foss-2020a-Python-3.8.2
  gmx ...