ray/release/rllib_tests/unit_gpu_tests/cluster.yaml

44 lines
1.2 KiB
YAML

cluster_name: ray-rllib-regression-tests
min_workers: 0
max_workers: 0
docker:
image: anyscale/ray-ml:latest-gpu
container_name: ray_container
pull_before_run: True
# Cloud-provider specific configuration.
provider:
type: aws
region: us-west-2
availability_zone: us-west-2a
cache_stopped_nodes: False
# How Ray will authenticate with newly launched nodes.
auth:
ssh_user: ubuntu
head_node:
InstanceType: p2.xlarge # Cheaper 1GPU K80 instance
# Set primary volume to 25 GiB
BlockDeviceMappings:
- DeviceName: /dev/sda1
Ebs:
VolumeSize: 100
# List of shell commands to run to set up nodes.
setup_commands:
- apt-get install -y libglib2.0-0 libcudnn7=7.6.5.32-1+cuda10.1 curl unzip gcc python3-dev
# Command to start ray on the head node. You don't need to change this.
head_start_ray_commands:
- ray stop
- ulimit -n 65536; OMP_NUM_THREADS=1 ray start --head --redis-port=6379 --object-manager-port=8076 --autoscaling-config=~/ray_bootstrap_config.yaml
# Command to start ray on worker nodes. You don't need to change this.
worker_start_ray_commands:
- ray stop
- ulimit -n 65536; OMP_NUM_THREADS=1 ray start --address=$RAY_HEAD_IP:6379 --object-manager-port=8076