How to setup jupyter Notebook on Google Cloud Platform

Step 1: Add a VM instance
On Google Cloud Platform Home page Click
Compute Engine → VM instances

Click Create button

firstly give your instance a name I write “notebook” for this instance. If you are an AMD fan select us-central1 (lowa) to be able to choose the N2D cpu series, for this project I really need some compute power so I choose
n2d-standard-4 which has 4 cores and 16 GB memory

on Boot disk part click the change button

for this project, I want to select Ubuntu 20.04 LTS and SSD persistent disk with 20GB volume.

on Firewall section select Allow HTTP traffic and Allow HTTPS traffic
You are ready to go, click the create button

After your VM instance setup finished you will be redirected to the VM instances page which looks like this

Step 2: Open port on firewall

I want to use port 60000 for this notebook so I will add a firewall rule

Click your instance

click network

Click Firewall

Click Create Firewall Rule

• give firewall rule a name. I write “notebook”
• targets: All instances in the network
• Source filter: Ip ranges
• Source IP ranges : 0.0.0.0/0
• second source filter : none
• Specified protocols and ports: tcp:60000

and click the create button. Now you can connect your notebook from tcp:60000.

Step 3: Install Jupyter Notebook

On the VM instances dashboard click the SSH button to connect your VM instance

now your browser will open another window which is your console

you need to enter all commands in this console let's start with an upgrade

sudo apt update && sudo apt upgrade -y

now visit this URL and find the latest release of Anaconda for Linux-x86_64 and right-click the link and copy it.

for now, Anaconda3–2021.11-Linux-x86_64.sh is the latest release. We use curl -O command to download the file

cd ~ && curl -O https://repo.anaconda.com/archive/Anaconda3-2021.11-Linux-x86_64.sh

this is a .sh file so we use the bash command to run it

bash Anaconda3-2021.11-Linux-x86_64.sh

hold ENTER till you see yes or no choice and write yes

hit the ENTER for the last time to confirm installation location and your installation will start

After installation finished it will ask to run initialize Anaconda3 say yes

! important : close current shell and reconnect with ssh. Its because jupyter command not recognized with current ssh session.

now we need to generate a jupyter notebook configuration file with this command

jupyter notebook --generate-config

After generating the config file command you will see the path of the jupyter_notebook_config.py file copy this path and edit it with nano file editor

nano /home/beyondtheclouds0001/.jupyter/jupyter_notebook_config.py

you can simply copy and paste these lines at the beginning of the configuration file

c.NotebookApp.token = ''
c.NotebookApp.ip = '*'
c.NotebookApp.open_browser = False
c.NotebookApp.port = 60000

control + x to exit and press Y to save changes then press enter.

you are ready to go. Just start jupyter notebook with this command

jupyter-notebook

Go to VM instances dashboard and copy your instance external IP address

open a new browser and enter your IP address and port number as it is

Step 4:How to run Jupiter Notebook as a service

When you close your ssh window, the notebook process will stop If you need this process constantly run on this virtual machine you need to convert this process to a Linux service.

Add a new file under the “/etc/systemd/system/” directory
this file name will be “jupyter.service”.

nano /etc/systemd/system/jupyter.service

change “beyondtheclouds0001” parts then copy-paste codes below to your notebook service file and save the file.

Note: in this example, my username on Linux is “beyondtheclouds0001”

where is my jupyter-notebook bin folder?
it is under the anaconda3 folder where you downloadAnaconda3–2021.11-Linux-x86_64.sh and run it. It auto-makes a folder in the same directory. In this example, I downloaded Anaconda to my home directory which is /home/beyondtheclouds0001/

[Unit]
Description=Jupyter Notebook
[Service]
Type=simple
PIDFile=/run/jupyter.pid
ExecStart=/home/beyondtheclouds0001/anaconda3/bin/jupyter-notebook ---config=/home/beyondtheclouds0001/.jupyter/jupyter_notebook_config.py
User=beyondtheclouds0001
Group=google-sudoers
WorkingDirectory=/home/beyondtheclouds0001/
Restart=always
RestartSec=10
[Install]
WantedBy=multi-user.target

so in the end, there will be a “/etc/systemd/system/jupyter.service” file

we need to reload daemon so linux system will see our new jupyter.service file

sudo systemctl daemon-reload

we need to make this service started when Linux makes a restart so run this command

sudo systemctl enable jupyter

start jupyter service

sudo systemctl start jupyter

check if jupyter service is running

sudo systemctl status jupyter

these commands make your notebook service as an initial service and when you start your virtual machine notebook service will start automatically.

Do not hesitate to ask questions. keep coding 😜