ly01325/IsaacSim-Automator

Tools for Deployment of Isaac Sim to Public Clouds

Isaac Sim Automator

• Installation
  • Installing Docker
  • Obtaining NGC API Key
  • Building the Container
• Usage
  • Tip: Running the Automator Commands
  • Deploying Isaac Sim
    • AWS
    • GCP
    • Azure
    • Alibaba Cloud
  • Connecting to Deployed Instances
  • Running Applications
    • Isaac Sim
    • Shell in Isaac Sim Container
    • Omniverse Isaac Gym
  • Pausing and Resuming
  • Uploading Data
  • Downloading Data
  • Destroying
• Known Issues
  • Apple Silicon Support

This tool automates deployment of Isaac Sim to public clouds.

Installation

Installing Docker

docker should be present on your system. Visit https://docs.docker.com/engine/install/ for installation instructions.

Obtaining NGC API Key

NGC API Key allows you to download docker images from https://ngc.nvidia.com/. Please prepare one or obtain it at https://ngc.nvidia.com/setup/api-key.

Building the Container

Please enter the following command in the project root directory to build the container:

./build

This will build the Isaac Sim Automator container and tag it as isa.

Usage

Tip: Running the Automator Commands

There are two ways to run the automator commands:

1. First enter the automator container and then run the command inside the container:

# enter the automator container
./run
# inside container:
./someconnad

2. Simply prepend the command with ./run like so:

./run ./somecommand <parameters>

for example:

./run ./deploy-aws
./run ./destroy my-deployment

Deploying Isaac Sim

AWS

Enabling Access Permissions

You need AmazonEC2FullAccess permissions enabled for your AWS user. You can enable those in Identity and Access Management (IAM) Section in AWS console like so:

1. Go to https://console.aws.amazon.com/iamv2/home#/home
2. Click "Access Management" > "Users" in the left menu
3. Search for your user name
4. Under "Permissions" tab click "Add permissions"
5. Choose "Attach existing policies directly"
6. Search for AmazonEC2FullAccess, check the box next to it, click "Next"
7. Click "Add permissions"

Getting Access Credentials

You will need _AWS Access Key_ and _AWS Secret Key_ for an existing account. You can obtain those in Identity and Access Management (IAM) Section in the AWS console.

If yoou have completed the above steps or already have your permissions and credentials set up, run the following command in the project root directory:

# enter the automator container
./run
# inside container:
./deploy-aws

Tip: Run ./deploy-aws --help to see more options.

GCP

# enter the automator container
./run
# inside container:
./deploy-gcp

Tip: Run ./deploy-gcp --help to see more options.

Azure

If You Have Single Subscription:

# enter the automator container
./run
# inside container:
./deploy-azure

If You Have Multiple Subscriptions:

 # enter the automator container
./run

# inside container:
az login # login
az account show --output table # list subscriptions
az account set --subscription "<subscription_name>"
./deploy-azure --no-login

Tip: Run ./deploy-azure --help to see more options.

Alibaba Cloud

Getting Access Credentials

You will need Access Key and Secret Key for an existing AliCloud account. You can obtain those in AccessKey Management section in the Alibaba Cloud console.

Once you have prepared the access credentials, run the following command in the project root directory:

# enter the automator container
./run
# inside container:
./deploy-alicloud

Tip: Run ./deploy-alicloud --help to see more options.

GPU-accelerated instances with NVIDIA A100, A10 and T4 GPUs are supported. You can find the complete list of instance types, availability and pricing at https://www.alibabacloud.com/help/en/ecs/user-guide/gpu-accelerated-compute-optimized-and-vgpu-accelerated-instance-families-1. Please note that vGPU instances are not supported.

Connecting to Deployed Instances

Deployed Isaac Sim instances can be accessed via:

• SSH
• noVNC (browser-based VNC client)
• NoMachine (remote desktop client)

Look for the connection instructions at the end of the deploymnt command output. Additionally, this info is saved in state/<deployment-name>/info.txt file.

You can view available arguments with --help switch for the start scripts, in most cases you wouldn't need to change the defaults.

Tip: You can use ./connect <deployment-name> helper command to connect to the deployed instance via ssh.

Running Applications

To use installed applications, connect to the deployed instance using noVNC or NoMachine. You can find the connection instructions at the end of the deployment command output. Additionally, this info is saved in state/<deployment-name>/info.txt file.

Isaac Sim

Isaac Sim will be automatically started when cloud VM is deployed. Alternatively you can click "Isaac Sim" icon on the desktop or run the following command in the terminal on the deployed instance or launch it from the terminal as follows:

~/Desktop/isaacsim.sh

Shell in Isaac Sim Container

To get a shell inside Isaac Sim container, click "Isaac Sim Shell" icon on the desktop. Alternatively you can run the following command in the terminal on the deployed instance:

~/Desktop/isaacsim-shell.sh

Omniverse Isaac Gym

Omniverse Isaac Gym Reinforcement Learning Environments for Isaac Sim ("Omniverse Isaac Gym") is pre-installed on the deployed Isaac instances.

To run Omniverse Isaac Gym click "Omni Isaac Gym" icon on the desktop or run the following command in the terminal:

~/Desktop/omni-isaac-gym-envs.sh

Default output directory (/OmniIsaacGymEnvs/omniisaacgymenvs/runs) in the OmniIsaacGymEnvs contaner will be linked to the default results directory (/home/ubuntu/results) on the deployed instance. You can download the contents of this directory to your local machine using ./download <deployment_name> command.

Pausing and Resuming

You can stop and re-start instances to save on cloud costs. To do so, run the following commands:

# enter the automator container
./run
# inside container:
./stop <deployment-name>
./start <deployment-name>

Currently, stop-start is only supported for Azure deployments, other clouds will be added soon.

Uploading Data

You can upload user data from uploads/ folder (in the project root) to the deployment by running the following command:

# enter the automator container
./run
# inside container:
./upload <deployment-name>

Data will be uploaded to /home/ubuntu/uploads directory by default to all deployed instances. You can change this by passing --remote-dir argument to the command. Run ./upload --help to see more options.

Downloading Data

You can download user data to results/ folder (in the project root) from deployed instances by running the following command:

# enter the automator container
./run
# inside container:
./download <deployment-name>

Data will be downloaded from /home/ubuntu/results directory by default. You can change this by passing --remote-dir argument to the command. Run ./download --help to see more options.

Destroying

To destroy a deployment, run the following command:

# enter the automator container
./run
# inside container:
./destroy <deployment-name>

You will be prompted to enter the deployment name to destroy.

Please note that information about the deployed cloud resources is stored in state/ directory. Do not delete this directory ever.

Known Issues

Apple Silicon Support

Issue:

Some packages in the application docker container are not compatible with ARM architecture of Apple Silicon (M1/M2/etc).

Workaround:

Build the ap container using Docker Desktop or standalone Docker installation with the --platform linux/x86_64 flag like so:

docker build -t isa --platform linux/x86_64 .