MoltenIron overview

MoltenIron maintains a pool of bare metal nodes.

Starting

Before starting the server for the first time, the createDB.py script must be run.

To start the server:

moltenirond-helper start

To stop the server:

moltenirond-helper stop

MoltenIron client

Use the molteniron client (molteniron) to communicate with the server. For usage information type:

molteniron -h

For usage of a specific command use:

molteniron [command] -h

MoltenIron commands

command	description
add	Add a node
allocate	Allocate a node
release	Release a node
get_field	Get a specific field in a node
set_field	Set a specific field with a value in a node
status	Return the status of every node
delete_db	Delete every database entry

Configuration of MoltenIron

Configuration of MoltenIron is specified in the file conf.yaml.

"B)" means that this configuration option is required for both the client and the server. "C)" means that it is required only for the client. "S)" means it is only required for the server.

usage	key	description
B)	mi_port	the port that the server uses to respond to commands.
C)	serverIP	The IP address of the server. This is only used by
	clients.
S)	maxTime	The maximum amount of time, in seconds, that a node
	is allowed to be allocated to a particular BM node.
S)	logdir	The path to the directory where the logs should be
	stored.
S)	maxLogDays	The amount of time, in days, to keep old logs.
S)	sqlUser	The username to use for the MI server. This user
	will automatically be generated when createDB.py is run.
S)	sqlPass	The password of sqlUser

Running testcases

The suite of testcases is checked by tox. But, before you can run tox, you need to change the local yaml configuration file to point to the log directory.

(LOG=$(pwd)/testenv/log; sed -i -r -e 's,^(logdir: )(.*)$,\1'${LOG}',' conf.yaml; rm -rf testenv/; tox -e testenv)

Running inside a Continuous Integration environment

During the creation of a job, add the following snippet of bash code:

# Setup MoltenIron and all necessary prerequisites.
# And then call the MI script to allocate a node.
(
  REPO_DIR=/opt/stack/new/third-party-ci-tools
  MI_CONF_DIR=/usr/local/etc/molteniron/
  MI_IP=10.1.2.3 # @TODO - Replace with your IP addr here!

  # Grab molteniron and install it
  git clone https://git.openstack.org/openstack/third-party-ci-tools ${REPO_DIR} || exit 1

  cd ${REPO_DIR}/nodepool/molteniron

  # @BUG Install prerequisite before running pip to install the requisites
  hash mysql_config || sudo apt install -y libmysqlclient-dev

  # Install the requisites for this package
  sudo pip install --upgrade --force-reinstall --requirement requirements.txt

  # Run the python package installation program
  sudo python setup.py install

  if [ -n "${MI_IP}" ]
  then
    # Set the molteniron server IP in the conf file
    sudo sed -i "s/127.0.0.1/${MI_IP}/g" ${MI_CONF_DIR}/conf.yaml
  fi

  export dsvm_uuid
  # NOTE: dsvm_uuid used in the following script, hence the -E
  sudo -E ${REPO_DIR}/nodepool/molteniron/utils/test_hook_configure_mi.sh
) || exit $?

and change the MI_IP environment variable to be your MoltenIron server!

During the destruction of a job, add the following snippet of bash code:

  DSVM_UUID="$(</etc/nodepool/uuid)"
  echo "Cleaning up resources associated with node: ${DSVM_UUID}"
  molteniron release ${DSVM_UUID}