# Minimal configuration for an automatically managed on-premise cluster.

# To use, run the script at ray/python/ray/autoscaler/local/coordinator_server.py:
# $ python coordinator_server.py --ips <list_of_node_ips> --host <HOST> --port <PORT>
# Copy the address from the output into the coordinator_address field.

# A unique identifier for the head node and workers of this cluster.
cluster_name: minimal-automatic

provider:
    type: local
    coordinator_address: COORDINATOR_HOST:COORDINATOR_PORT

# The minimum number of workers nodes to add to the Ray cluster in addition to the head
# node. This number should be >= 0.
# Set to 0 by default.
min_workers: 0

# The maximum number of worker nodes to add to the Ray cluster in addition to the head node.
# This takes precedence over min_workers.
# Required for automatically managed clusters.
max_workers: 2

# How Ray will authenticate with newly launched nodes.
auth:
    ssh_user: YOUR_USERNAME
    # Optional if an ssh private key is necessary to ssh to the cluster.
    # ssh_private_key: ~/.ssh/id_rsa

# The above configuration assumes Ray is installed on your on-prem cluster.
# If Ray is not already installed on your cluster, you can use setup
# commands to install it.
# For the latest Python 3.7 Linux wheels:
# setup_commands:
  # - if [ $(which ray) ]; then pip uninstall ray -y; fi
  # - pip install -U "ray[default] @ https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-3.0.0.dev0-cp37-cp37m-manylinux2014_x86_64.whl"

# Defaults are empty to avoid any surprise changes to on-prem cluster's state.
# Refer to example yamls for examples of ray installation in setup commands.
initialization_commands: []
setup_commands: []
head_setup_commands: []
worker_setup_commands: []

available_node_types: {}
head_node_type: {}

head_start_ray_commands: []
worker_start_ray_commands: []

file_mounts: {}
cluster_synced_files: []
file_mounts_sync_continuously: false
rsync_exclude: []
rsync_filter: []