gcp_setup.md

Setting up a VM on GCP

Step 1: Login to GCP console and create project

Name: mlops-demo

Step 2: Create an instance using Compute Engine

Configuration:

Name: mlops-zoomcamp
Region: asia-northeast3 (Seoul)
Zone: asia-northeast3-a
Series: N2
Machine Type: n2-standard-4
Boot disk:
Image: Ubuntu 22.04 LTS (30 GB)
Firewall: Check both
Allow HTTP traffic
Allow HTTPS traffic

Step 3: Assign static IP

View Network details > IP addresses > Reserve external static IP address

Step 4: Enable SSH

Generate and copy key

ssh-keygen -t rsa -f ~/.ssh/mlops-zoomcamp -C pytholic
cat ~/.ssh/mlops-zoomcamp.pub

On GCP console, enter the copied public key

Compute Engine*> Metadata > SSH Key*

Step 5: SSH into instance

ssh -i ~/.ssh/mlops-zoomcamp pytholic@<External VM IP>

Optional, Add host to hostname file for easy login

Enter the following into ~/.ssh/config

vim ~/.ssh/config

Host gcp-mlops-zoomcamp
    HostName 34.64.207.252 # VM Public IP
    User pytholic # VM user
    IdentityFile ~/.ssh/mlops-zoomcamp # Private SSH key file
    StrictHostKeyChecking no

ssh gcp-mlops-zoomcamp

Step 6: Install dependencies

• Anaconda

cd ~
wget https://repo.anaconda.com/archive/Anaconda3-2022.05-Linux-x86_64.sh
bash Anaconda3-2022.05-Linux-x86_64.sh

• Docker

sudo apt update
sudo apt install docker.io

• Docker-Compose

sudo apt install docker-compose

Add current user to docker group

sudo usermod -aG docker $USER

Log out and log in again.

• Verify installations

which python
# /home/pytholic/anaconda3/bin/python

which docker
# /usr/bin/docker

which docker-compose
# /usr/bin/docker-compose

docker run hello-world

Step 7: Connect with VS Code

• Install Remote SSH extension
• Connect to host
• Select gcp-mlops-zoomcamp

Step 8: Connect Jupyter Notebook

• Run notebook in remote terminal jupyter notebook
• Use port forwarding
• In VS Code
  • Go to ports tab and forward 8888

• Go to localhost:8888
• Copy token from terminal and set a new password