ibm-vpc-ray-connector 1.0.0b70

IBM VPC connector for Ray

ibm-vpc-ray-connector enables Ray cluster to be deployed over IBM VPC infrastructure

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

1. Install ibm-vpc-ray-connector on your machine

pip install ibm-vpc-ray-connector

2. Configure ibm vpc

   • Generate API KEY

   • Create/Update security group to have SSH, Redis and Ray Dashboard ports open: 22, 8265 and 6379

3. 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
   

4. 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