The bootmachine-admin start
command simply copies two files
to the current working directory. A standard Fabric fabfile.py
and
a settings.py
for which you need to customize. Additionally it
copies over the configuration
folder containing the initial
states/recipes to configure the servers.
After customizing your settings, all it takes to convert bare metal servers from aluminium into rhodium is one simple Fabric command:
$ fab bootmachine
Internally this does two things. First provider.bootem
checks if
there are any non-booted servers listed in the
settings.PROVIDER_BACKENDS
. If provider.boot
finds a
non-booted server, it will boot it in parallel. In the meantime
provider.bootem
queries the provider to ensure that all servers are
ACTIVE
before continuing.
Second, after all servers are found to be active, the bootstrap
method is called to check if there are any servers which have not yet
been bootstrapped. These servers are then bootstrapped in parallel.
Note
These commands can also be run separately:
$ fab boot $ fab each bootstrap_distro $ fab each bootstrap_configurator $ fab master configure
After provisioning is complete you can manually login, with the user credentials and port as defined in your settings.py, to the machine using the following format:
$ ssh -p {port} {username}@{ip}
To list the details for your new machines, including ip addresses:
$ fab provider
Or if you already have openstack-compute
or python-novaclient
installed, you could just as easily:
$ openstack-compute list or $ nova list
All available commands can be seen by typing:
$ fab -l # which is short for fab --list