The installation of the local k8s cluster with install_kubeadm_cluster.sh
raised an error message when deleting the namespace "osm". That NS is
deleted as a previous step to the deployment of OSM pods, just to make sure
that we start from a clean NS.

In principle, the error from kubectl should appear always in clean
installations. For that reason, the output of the command is redirected
to /dev/null to avoid confusion for the end user.

Change-Id: I65b90675194848f3968a9c192427224652e3087f
Signed-off-by: garciadeblas <gerardo.garciadeblas@telefonica.com>

Change-Id: Ibed733d04770f00789f4f0c2ed1faf77ffdb5143
Signed-off-by: garciadeblas <gerardo.garciadeblas@telefonica.com>

Change-Id: Ie83a4c9284ea63b40c37975da261634f930b6317
Signed-off-by: garciadeblas <gerardo.garciadeblas@telefonica.com>

This change covers:

- Feature 10892. Installation of OSM on top of Ubuntu20.04. Changes
  are mostly in full_install_osm.sh and are related to the use of new
  versions of kubeadm and docker-ce. In addition, changes in Jenkins
  groovy files have been done to indicate the base image to be used,
  either 18.04 or 20.04.
- Feature 10893. Better tracking of installation. The code for tracking
  in in common/track. There is a function track that it is called in
  the different steps of the installation.
- Feature 8460: Cleanup old code in full_install_osm.sh. The script
  full_install_osm.sh has been split in different scripts performing
  specific tasks, thus simplifying the installer: install_docker_ce.sh,
  install_juju.sh and install_kubeadm_cluster.sh.

Change-Id: I1e388ec56285337eaf34f68470aa5a9b23ff45ff
Signed-off-by: garciadeblas <gerardo.garciadeblas@telefonica.com>