1+ # An unique identifier for the head node and workers of this cluster.
2+ cluster_name: default
3+
4+ # The maximum number of workers nodes to launch in addition to the head
5+ # node.
6+ max_workers: 2
7+
8+ # The autoscaler will scale up the cluster faster with higher upscaling speed.
9+ # E.g., if the task requires adding more nodes then autoscaler will gradually
10+ # scale up the cluster in chunks of upscaling_speed*currently_running_nodes.
11+ # This number should be > 0.
12+ upscaling_speed: 1.0
13+
14+ # This executes all commands on all nodes in the docker container,
15+ # and opens all the necessary ports to support the Ray cluster.
16+ # Empty string means disabled.
17+ docker:
18+ image: "orfs-autotuner:latest"
19+ container_name: "ray_container"
20+ # If true, pulls latest version of image. Otherwise, `docker run` will only pull the image
21+ # if no cached version is present.
22+ pull_before_run: false
23+
24+ # If a node is idle for this many minutes, it will be removed.
25+ idle_timeout_minutes: 5
26+
27+ # Cloud-provider specific configuration.
28+ provider:
29+ type: gcp
30+ region: us-west1
31+ availability_zone: us-west1-a
32+ project_id: foss-fpga-tools-ext-openroad
33+
34+ # How Ray will authenticate with newly launched nodes.
35+ auth:
36+ ssh_user: ubuntu
37+
38+ available_node_types:
39+ ray_head_default:
40+ resources: {"CPU": 2}
41+ node_config:
42+ machineType: n1-standard-2
43+ disks:
44+ - boot: true
45+ autoDelete: true
46+ type: PERSISTENT
47+ initializeParams:
48+ diskSizeGb: 50
49+ sourceImage: projects/deeplearning-platform-release/global/images/common-cpu-v20240922
50+ ray_worker_small:
51+ # The minimum number of worker nodes of this type to launch.
52+ # This number should be >= 0.
53+ min_workers: 1
54+ # The maximum number of worker nodes of this type to launch.
55+ # This takes precedence over min_workers.
56+ max_workers: 2
57+ # The resources provided by this node type.
58+ resources: {"CPU": 2}
59+ node_config:
60+ machineType: n1-standard-2
61+ disks:
62+ - boot: true
63+ autoDelete: true
64+ type: PERSISTENT
65+ initializeParams:
66+ diskSizeGb: 50
67+ sourceImage: projects/deeplearning-platform-release/global/images/common-cpu-v20240922
68+ # scheduling:
69+ # - preemptible: true
70+ # Un-Comment this to launch workers with the Service Account of the Head Node
71+ # serviceAccounts:
72+ # - email: ray-autoscaler-sa-v1@<project_id>.iam.gserviceaccount.com
73+ # scopes:
74+ # - https://www.googleapis.com/auth/cloud-platform
75+
76+ # Specify the node type of the head node (as configured above).
77+ head_node_type: ray_head_default
78+
79+ # Files or directories to copy to the head and worker nodes. The format is a
80+ # dictionary from REMOTE_PATH: LOCAL_PATH, e.g.
81+ file_mounts: {
82+ # "/path1/on/remote/machine": "/path1/on/local/machine",
83+ # "/path2/on/remote/machine": "/path2/on/local/machine",
84+ }
85+
86+ # Files or directories to copy from the head node to the worker nodes. The format is a
87+ # list of paths. The same path on the head node will be copied to the worker node.
88+ # This behavior is a subset of the file_mounts behavior. In the vast majority of cases
89+ # you should just use file_mounts. Only use this if you know what you're doing!
90+ cluster_synced_files: []
91+
92+ # Whether changes to directories in file_mounts or cluster_synced_files in the head node
93+ # should sync to the worker node continuously
94+ file_mounts_sync_continuously: False
95+
96+ # Patterns for files to exclude when running rsync up or rsync down
97+ rsync_exclude:
98+ - "**/.git"
99+ - "**/.git/**"
100+
101+ # Pattern files to use for filtering out files when running rsync up or rsync down. The file is searched for
102+ # in the source directory and recursively through all subdirectories. For example, if .gitignore is provided
103+ # as a value, the behavior will match git's behavior for finding and using .gitignore files.
104+ rsync_filter:
105+ - ".gitignore"
106+
107+ initialization_commands:
108+ - curl -fsSL https://get.docker.com -o get-docker.sh
109+ - sudo sh get-docker.sh
110+ - sudo usermod -aG docker $USER
111+ - sudo systemctl restart docker -f
112+
113+ # List of shell commands to run to set up nodes.
114+ setup_commands: []
115+ # Note: if you're developing Ray, you probably want to create a Docker image that
116+ # has your Ray repo pre-cloned. Then, you can replace the pip installs
117+ # below with a git checkout <your_sha> (and possibly a recompile).
118+ # To run the nightly version of ray (as opposed to the latest), either use a rayproject docker image
119+ # that has the "nightly" (e.g. "rayproject/ray-ml:nightly-gpu") or uncomment the following line:
120+ # - pip install -U "ray[default] @ https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-3.0.0.dev0-cp37-cp37m-manylinux2014_x86_64.whl"
121+
122+
123+ # Custom commands that will be run on the head node after common setup.
124+ head_setup_commands:
125+ - pip install google-api-python-client==1.7.8
126+
127+ # Custom commands that will be run on worker nodes after common setup.
128+ worker_setup_commands: []
129+
130+ # Command to start ray on the head node. You don't need to change this.
131+ head_start_ray_commands:
132+ - ray stop
133+ - >-
134+ ray start
135+ --head
136+ --port=6379
137+ --object-manager-port=8076
138+ --autoscaling-config=~/ray_bootstrap_config.yaml
139+
140+ # Command to start ray on worker nodes. You don't need to change this.
141+ worker_start_ray_commands:
142+ - ray stop
143+ - >-
144+ ray start
145+ --address=$RAY_HEAD_IP:6379
146+ --object-manager-port=8076
0 commit comments