Use of python virtual environment, e.g. virtualenv, is greatly encouraged to avoid installing Python packages globally, which could break system tools or other projects
- Install ibm-vpc-ray-connector on your machine
pip install ibm-vpc-ray-connector
-
Configure ibm vpc
-
Create/Update security group to have SSH, Redis and Ray Dashboard ports open: 22, 8265 and 6379
-
Create cluster config file
- Use interactive
ibm-ray-config
config tool to generate cluster.yaml configuration file
pip install ibm-ray-config ibm-ray-config -o cluster.yaml
- Use interactive
- Use generated file to bring ray cluster up, e.g
ray up cluster.yaml
.
- After finished, find cluster head node and worker nodes ips:
ray get-head-ip cluster.yaml
ray get-worker-ips cluster.yaml
- To get status of the cluster
ray status --address PUBLIC_HEAD_IP:6379
-
Use browser to open ray dashboard on PUBLIC_HEAD_IP:8265. Alternatively use
ray dashboard
to forward ray cluster dashboard to your localhost. -
Submit example task
ray submit cluster.yaml templates/example.py
Logs for the node_provider can be found under /tmp/connector_logs/
.
Logs of all levels will be written to connector_logs
.
The default log level for console output is INFO
.