Fixes#1
The Vagrant file being used is a copy from
`https://github.com/stationgroup/vagrant-labs/tree/master/imperialspeculate`.
Two roles are being used: debian-upgrade (an upstream Galaxy role) and
freebsd-upgrade (a small role based on what was proposed in the comments of #1
and extended with proper support for check mode.)
The upgrade process is contained in the playbook os_upgrade.yml, which will
automatically create proper groups for Ubuntu and FreeBSD hosts. If unneeded,
this first play can be left out, and the target hosts: in the second play can be
replaced by the relevant groups you have in the inventory (e.g. ec2 tags.)
A local ansible.cfg is defined, and needed for these scripts to run out of the
box. This implies that all ansible commands must be run from the
`ansible-experiments/package_updates` folder.
A small script `setup-requirements` is provided, that initializes everything,
to be executed after the vagrant boxes came online. It will generate an
ssh-config for said vagrant boxes, download roles from galaxy, and make a
base-install for the hosts (installing python dependencies, ansible itself
on ubuntu1, a deploying an ssh key to all nodes to be used from the vagrant box
`ubuntu1`, as ansible controller machine.)
When deploying and setting up from the machine where vagrant runs, you need to
add some extra arguments:
`--ssh-extra-args "-F ./vagrant-ssh-config" --inventory hosts-vagrant`
to ansible execution. These are not necessary once running ansible from
`ubuntu1`.
The vagrant setup seems to have a provisioning bug, that kicks in with the
latest 18.04 Ubuntu. The FreeBSD boxes also experience a provisioning problem,
with the same result: the second, private network interface does not get
configured. As these interfaces are used to run ansible from `ubuntu1`, I could
not fully test the scripts from there.
