*NEW* Public IPs, ports, and SSH into running jobs 🌟

tutorial
gradient
announcement

#1

Hey there, we just pushed better support for public IP addresses and ports for your Gradient jobs. Easily run TensorBoard, SSH in to a running job, or even run a Flask web service!

Learn more at the heldesk here: https://paperspace.zendesk.com/hc/en-us/articles/360003412574-4-Public-IPs-and-port-forwarding

and here’s a sample workspace to SSH in to a running Gradient job:

Let us know what you think!


#2

I tried using this example, and when i SSH into the public ip of the running job, it seems to stall and then gets this error: ssh: connect to host 35.230.100.188 port 8888: Operation timed out

please fix!!!

note: I only want to ssh into a running job so i can see the cpu/gpu and memory resource metrics


#3

@gemadang This worked for me. Here’s what I ran:

# Clone the repo
git clone https://github.com/Paperspace/gradient-ssh-test

# Navigate into the cloned repo directory
cd gradient-ssh-test/

# Initialize the project in Gradient
paperspace project init

# Run the job
paperspace jobs create --machineType P100 --container Test-Container --ports 8888:22 --command './run.sh'

# Wait for the Job to enter the running state
# From a new terminal, SSH into your job using the public IP listed in the Job logs/console
ssh [email protected] -p 8888

# The password is listed in Github: [email protected]