ray/release/rllib_tests/autoscaler-cluster.yaml

59 lines
1.9 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
run_options: ["--ulimit nofile=1045876"]
# 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: p3.16xlarge
ImageId: ami-0a2363a9cff180a64 # Custom ami
# Set primary volume to 25 GiB
BlockDeviceMappings:
- DeviceName: /dev/sda1
Ebs:
VolumeSize: 100
worker_nodes:
InstanceType: p3.16xlarge
ImageId: ami-0a2363a9cff180a64 # Custom ami
# 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
- pip install terminado
- pip install torch==1.6 torchvision
- pip install boto3==1.4.8 cython==0.29.0
- "pip install https://s3-us-west-2.amazonaws.com/ray-wheels/releases/1.1.0/f591f6c1c8fa14af6df2adfdcf3255c59dff12b1/ray-1.1.0.dev0-cp37-cp37m-manylinux2014_x86_64.whl"
- git clone https://github.com/ray-project/ray.git ray-cp
- pip install -r ./ray-cp/release/rllib_tests/regression_tests/requirements.txt
# Command to start ray on the head node. You don't need to change this.
head_start_ray_commands:
- ray stop
- OMP_NUM_THREADS=1 ray start --head --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
- OMP_NUM_THREADS=1 ray start --address=$RAY_HEAD_IP:6379 --object-manager-port=8076