Single Node

If you requested single node resources via OOD, ARE or Gadi PBS job, you can simply call ray.init() within your Jupyter Notebook or python script to start a Ray runtime on the current host.

You can view the resources available to the Ray via function ray.cluster_resources().

import ray
ray.init()
print(ray.cluster_resources())

The above commands may print some messages as below indicating there are 16 CPU core, 29GB memory and 1 node available for the current local Ray cluster.

{'memory': 28808935835.0, 'object_store_memory': 14404467916.0, 'CPU': 16.0, 'node:10.0.128.152': 1.0}

Multiple Nodes

If you need a larger scale of Ray cluster crossing multiple nodes, you can start a pre-defined Ray cluster and then connect it in the Jupyter notebook or python script.

The easiest way to set up the pre-defined Ray cluster is to utilize dk92 module "gadi_jupyterlab".  

Launching a predefined Ray cluster

Gadi

In your PBS job script, you should load the gadi_jupyterlab module together with NCI-data-analysis/2022.06. Then you need to run a script called "jupyter.ini.sh -R" to set up the predefined Ray cluster. It will start a Ray worker on each CPU core. You can also specify the number of Ray workers by following "-p" flag. For example, in a job requesting 96 cores ( 2 nodes), you can set up a pre-defined Ray cluster with 12 Ray workers per node, and 24 Ray workers in total via the following command:

jupyter.ini.sh -R -p 12

An example of a full job script is given below

#!/bin/bash
#PBS -P fp0
#PBS -q normal
#PBS -lwd,walltime=10:00:00,ncpus=96,mem=192GB,jobfs=400GB,storage=gdata/dk92+gdata/z00+scratch/fp0+gdata/fp0

module purge
module use /g/data/dk92/apps/Modules/modulefiles
module load NCI-data-analysis/22.06 gadi_jupyterlab/22.05

jupyter.ini.sh -R   # set up a Ray cluster with 48 Ray workers per node,96 total Ray workers, 1 thread per Ray worker.

# jupyter.ini.sh -R -p 12 # or set up a Ray cluster with 12 Ray workers per node, 24 total Ray workers, 1 threads per Ray worker.

python script.py

In the "script.py", you can connect to the predefined Ray cluster by calling ray.init() and specify the address flag as "auto".

import ray
ray.init(address="auto")
print(ray.cluster_resources())

The above script will print the following messages

{'object_store_memory': 114296048025.0, 'CPU': 96.0, 'memory': 256690778727.0, 'node:10.6.48.66': 1.0, 'node:10.6.48.67': 1.0}

ARE

First of all, you need to request multiple nodes in ARE JupyterLab session interface and specify the proper storage projects.

Then click "advanced options", put "/g/data/dk92/apps/Modules/modulefiles" in "Module directories" and load both NCI-data-analysis/22.06 and gadi-jupyterlab/2022.06 modules in "Module" field. In the "Pre-script" field, fill in the command "jupyterlab.ini.sh -R" to set up the predefined Ray cluster.  

Click "Open JupyterLab" button to open the JupyterLab session as soon it is highlighted.

In the Jupyter notebook, using the following lines to connect the predefined Ray cluster and print the available resources information. 

You can see 96 CPU Cores and two nodes are used by the cluster as expected..

Monitoring Ray status

You can easily monitor Ray status via the command "ray status".  Open a CLI terminal in either JupyterLab session or a Gadi PBS interactive job and type in the following command

 $ watch ray status

The Ray status will be kept updating every 2 seconds

Every 2.0s: ray status gadi-cpu-clx-1146.gadi.nci.org.au: Mon May 23 15:27:27 2022

======== Autoscaler status: 2022-05-23 15:27:26.840051 ========
Node status
---------------------------------------------------------------
Healthy:
1 node_ab2c1ef9316a7ee01bae8cd9d087f5e5dfbe3c8c254cf0e8752be0b1
1 node_b2f22752fc2b329e68649a81ba3c26f2d4f6080fa822dda0121d5514
Pending:
(no pending nodes)
Recent failures:
(no failures)

Resources
---------------------------------------------------------------
Usage:
96.0/96.0 CPU
0.00/239.062 GiB memory
23.70/106.446 GiB object_store_memory

Demands:
{'CPU': 1.0}: 225+ pending tasks/actors