How do I deploy Slurm on Railway on Railway?

You can deploy Slurm on Railway on Railway by clicking the "Deploy Now" button on this page. Railway will automatically set up all the necessary services and configurations for you.

What are the system requirements?

Railway handles all the infrastructure requirements. You only need a web browser to deploy and manage your application.

Is this template free to use?

Yes, this template is free to use within your existing Railway account.

Can I customize this template?

Yes, you can fully customize this template. After deployment, you will have full control of what you've deployed in your Railway project.

How do I get support for this template?

You can get support through our community forum: slurm-on-railway-603feb4b, or click "Get template help" for this template for more assistance.

Slurm on Railway

Why? Why not. POC of running Slurm controller + container-local worker nodes on Railway. Only local dependencies are railway CLI and docker.

Deployment

Deploy with railway template - be patient, the build takes about 10 minutes but subsequent deployments will be faster.
Create railway.env file with your project info:

# After deploying the template, you can get these from the URL:
https://railway.com/project/$PROJECT_ID/service/$SERVICE_ID?environmentId=$ENVIRONMENT_ID
export RAILWAY_PROJECT_ID=xxx
export RAILWAY_ENVIRONMENT_ID=xxx
export RAILWAY_SERVICE_ID=xxx

From the Settings tab of your project, get your public domain and port - e.g. interchange.proxy.rlwy.net:59019
Auth your Railway CLI with railway login
Run commands using the client.sh wrapper:

chmod +x client.sh
./client.sh  

# example
./client.sh interchange.proxy.rlwy.net:59019 scontrol ping -vvvv
Using Project: xxx
Using Environment: xxx
Using Service: xxx
--- Building Local Slurm Image ---
sha256:639351c1520234413d42a3df8b0230e3a04e317af4a1e305bbc35e775e750759
--- Syncing with Railway ---
Warning: Received unknown message type: stand_by
Remote hostname detected: 8f524205061f
Warning: Received unknown message type: stand_by
--- Launching Client Container ---
scontrol: debug2: _sack_connect: connected to /run/slurm/sack.socket
Slurmctld(primary) at 8f524205061f is UP

Client script

Builds a local Docker image slurm-railway.
Fetches the authentication key (created at build-time) and hostname (set by Railway at runtime) from Railway
Launch a background sackd (Slurm Auth and Cred Kiosk) daemon inside the container to handle the auth/slurm handshake.
Run given command / drop into bash

Deploy and Host

One-click deploy.

About Hosting

One-click deploy

Why Deploy

Play with Slurm :)

Common Use Cases

Play with Slurm :)

Dependencies for

None

Deployment Dependencies

None