Lab Setup on Google Cloud (GCP)¶

If your laptop has limited RAM/CPU resources, or you want a 24/7 dedicated cloud environment accessible from any machine —
This guide walks you through building an Ubuntu instance on Google Cloud Compute Engine, connecting securely from your laptop, and deploying Docker, Containerlab, and Arista cEOS end-to-end.

🧠 Architecture Overview¶

graph LR
    LAPTOP["💻 Laptop<br/>(Mac / Windows / Linux)"] -->|"SSH Connection<br/>(gcloud / Native Terminal / VS Code)"| VM["☁️ GCP Compute Engine VM<br/>(Ubuntu 24.04 LTS · 8 vCPU · 32GB RAM)"]
    VM --> DKR["🐳 Docker + Containerlab"]
    DKR --> FABRIC["⚡ Arista cEOS Fabric<br/>(Spines, Leafs, Hosts)"]

    classDef client fill:#1e293b,stroke:#38bdf8,color:#f8fafc,stroke-width:2px;
    classDef cloud fill:#0f172a,stroke:#a855f7,color:#f8fafc,stroke-width:2px;
    classDef nos fill:#064e3b,stroke:#34d399,color:#f8fafc,stroke-width:2px;
    class LAPTOP client; class VM cloud; class DKR,FABRIC nos;

No heavy virtualization software runs on your laptop; everything executes seamlessly inside the Google Cloud VM.

1. Prerequisites¶

Google Cloud Platform (GCP) Account — With billing enabled (cloud.google.com). Free trial credits ($300) work perfectly.
GCP Project ID — A dedicated project created for your labs.
Connecting Tool —
gcloud CLI installed on your laptop (instructions below), or
In-browser Google Cloud Shell (requires zero local installation).

2. Create the GCP VM Instance¶

You can build the VM using either the Web Console (GUI) or the gcloud CLI:

Option A · Google Cloud Console (Web GUI)¶

Open console.cloud.google.com and select your project.
Navigate to Navigation Menu (☰) → Compute Engine → VM instances.
Click Create Instance.
Configure the following parameters:
Name: netforge-lab
Region & Zone: Choose the closest region (e.g. asia-southeast1-b for Singapore, us-central1-a for US).
Machine Configuration:
- Series: E2 (cost-effective) or N2 (high performance).
- Machine type: e2-standard-8 (8 vCPU, 32 GB RAM) — comfortably runs 4 to 8 cEOS nodes.
Boot Disk:
- Click Change.
- Operating System: Ubuntu
- Version: Ubuntu 24.04 LTS (x86/64, amd64)
- Boot disk type: Balanced persistent disk
- Size: 100 GB (sized for container images and lab artifacts).
- Click Select.
Firewall: Check Allow HTTP traffic and Allow HTTPS traffic.
Click Create. Within 30–45 seconds, your VM is up and running with an external IP.

Option B · `gcloud` CLI (Command Line)¶

Run this single command from your laptop terminal or Cloud Shell:

# 1. Authenticate and configure defaults
gcloud auth login
gcloud config set project YOUR_PROJECT_ID
gcloud config set compute/zone asia-southeast1-b
gcloud config set compute/region asia-southeast1

# 2. Provision the VM
gcloud compute instances create netforge-lab \
  --zone=asia-southeast1-b \
  --machine-type=e2-standard-8 \
  --image-family=ubuntu-2404-lts-amd64 \
  --image-project=ubuntu-os-cloud \
  --boot-disk-size=100GB \
  --boot-disk-type=pd-balanced \
  --tags=netforge-lab

(If you plan to run nested QEMU/KVM nodes in the future, use the N2 series with --enable-nested-virtualization).

3. Connect from Laptop to GCP VM (4 Methods)¶

Method 1 · `gcloud compute ssh` (Recommended & Easiest)¶

From your laptop terminal:

gcloud compute ssh netforge-lab --zone=asia-southeast1-b

💡 Why it's best: gcloud automatically generates an SSH keypair, pushes your public key to the project metadata, and connects you directly to the VM shell without manual key management.

Method 2 · Native Terminal SSH using Standard SSH Keys¶

To use standard ssh from your macOS/Linux/WSL terminal:

Generate an SSH key on your laptop (if you haven't already):
```
ssh-keygen -t ed25519 -C "your-email@gmail.com"
```
Copy your public key:
```
cat ~/.ssh/id_ed25519.pub
```
In the GCP Console → Compute Engine → VM instances → Click netforge-lab → Click Edit → Under SSH Keys, click Add Item, paste the key, and click Save.

Retrieve the VM's external IP:

gcloud compute instances describe netforge-lab --format='get(networkInterfaces[0].accessConfigs[0].natIP)'

Connect directly:

ssh -i ~/.ssh/id_ed25519 <username>@<EXTERNAL_IP>

Method 3 · Cloud Console In-Browser SSH (1-Click)¶

In the GCP Console under Compute Engine → VM instances, click the SSH button next to netforge-lab. A full terminal window opens directly in your browser.

Method 4 · VS Code Remote-SSH¶

Edit configs and browse files with a full IDE experience:

Install the Remote - SSH extension in VS Code.

In your laptop's ~/.ssh/config, add:

Host netforge-gcp
    HostName <VM_EXTERNAL_IP>
    User <YOUR_USERNAME>
    IdentityFile ~/.ssh/id_ed25519

Click the Remote Window icon (><) in VS Code bottom-left → Connect to Host → netforge-gcp.

4. Install Docker, Containerlab, and cEOS on the GCP VM¶

Inside the GCP VM shell, install the required software:

4.1 Install Docker¶

curl -fsSL https://get.docker.com | sudo sh
sudo usermod -aG docker "$USER"
newgrp docker
docker run --rm hello-world     # Verifies: "Hello from Docker!"

4.2 Install Containerlab¶

bash -c "$(curl -sL https://get.containerlab.dev)"
containerlab version

4.3 Clone NetForge Labs¶

cd ~
git clone https://github.com/etherhtun/netforge-labs.git
cd netforge-labs

4.4 Import the Arista cEOS Image¶

Download cEOS64-lab-4.32.0F.tar.xz from arista.com to your laptop.

Upload the file from your laptop to the GCP VM using gcloud compute scp:

# Run on your laptop:
gcloud compute scp ~/Downloads/cEOS64-lab-4.32.0F.tar.xz netforge-lab:~/ --zone=asia-southeast1-b

Import into Docker inside the VM:
```
# Run on the GCP VM:
docker import cEOS64-lab-4.32.0F.tar.xz ceos:4.32.0F
docker images | grep ceos
```
(Because the GCP VM runs native x86_64 architecture, cEOS boots with zero emulation overhead!)

5. Deploy Your First Lab on GCP¶

cd ~/netforge-labs/labs/wan-edge-lab
sudo containerlab deploy -t topology.clab.yml

Inspect the running topology:

sudo containerlab inspect -t topology.clab.yml

Log in to a switch CLI:

docker exec -it clab-wan-edge-lab-edge1 Cli

6. Access Web Dashboards (Grafana / Prometheus / Web Graph)¶

To access web interfaces like Grafana (port 3000) or Containerlab Graph (port 50080) from your laptop browser:

Option 1 · GCP Firewall Rule:¶

gcloud compute firewall-rules create allow-netforge-web \
  --allow tcp:3000,tcp:9090,tcp:50080 \
  --target-tags=netforge-lab

(Access via http://<VM_EXTERNAL_IP>:3000).

Option 2 · SSH Port Forwarding (Most Secure):¶

ssh -L 3000:localhost:3000 -L 50080:localhost:50080 user@<VM_EXTERNAL_IP>

(Access securely via http://localhost:3000 on your laptop browser).

7. 💰 Save Costs: Stop the VM When Done¶

GCP charges for compute only while the instance is RUNNING. When finished for the day, shut it down:

# From inside the VM:
sudo poweroff

Or from your laptop:

gcloud compute instances stop netforge-lab --zone=asia-southeast1-b

To resume later:

gcloud compute instances start netforge-lab --zone=asia-southeast1-b

💡 Cost note: While stopped, compute billing drops to $0. You only pay standard persistent disk storage (~a few dollars/month for 100 GB). All your lab files, Docker images, and configurations are preserved.

8. Troubleshooting¶

Symptom	Cause	Solution
`Permission denied (publickey)` on SSH	Key not registered in GCP metadata	Use `gcloud compute ssh` or re-add public key under VM Edit → SSH Keys
`docker: command not found`	Docker not installed or group not active	Run `newgrp docker` or log out and back in
Out of memory when booting cEOS	VM has insufficient RAM	Stop the VM and upgrade machine type to `e2-standard-8` (32 GB)

🚀 Next Steps:¶

To learn the essential commands and topology structure for daily lab work, proceed to Page 3 · Docker & Containerlab CLI Guide →.