Introduction to Google Cloud Shell | Generated by AI

Home 2026.01 English 中文

---

Question: Comprehensive guide to Cloud Shell in Google Cloud

Answer:

Cloud Shell is a free, browser-based shell environment provided by Google Cloud that allows you to manage Google Cloud resources directly from the browser without installing any software locally.

What is Google Cloud Shell?

Cloud Shell provides a temporary Debian-based Linux VM (Compute Engine instance) pre-configured with:

• Google Cloud SDK (gcloud, gsutil, bq, etc.) — always up-to-date
• Popular tools: docker, terraform, kubectl, helm, git, node, python, go, java, etc.
• Authentication already handled (your Google account)
• 5 GB of persistent home directory storage (/home/<your_username>)
• Web-based code editor (Cloud Shell Editor — VS Code–like interface)
• Gemini CLI (AI assistance in terminal)

It is ideal for quick experiments, learning GCP, running gcloud commands, managing infrastructure as code, or light development.

How to Launch Cloud Shell

1. Go to the Google Cloud Console → https://console.cloud.google.com
2. Click the Cloud Shell icon (terminal symbol) in the top-right toolbar
   → or use direct link: https://shell.cloud.google.com
3. First time → it provisions your machine (takes ~10–30 seconds)
4. You get a Bash terminal at the bottom of the screen (can be popped out to new tab)

Alternative launch methods:

• cloudshell button in many GCP docs pages (“Open in Cloud Shell”)
• Direct deep-link with repo cloning: https://shell.cloud.google.com/?show=terminal&cloudshell_git_repo=<url>

Persistent Storage (5 GB Home Directory)

Only /home/$USER is persistent across sessions.

Everything else (/tmp, /usr/local, etc.) is ephemeral — deleted when session ends.

Common persistent locations:

• ~/ (home)
• ~/.bashrc, ~/.zshrc, ~/.gitconfig
• ~/projects/, ~/bin/, etc.

Tip: Store larger files in Cloud Storage buckets and mount them if needed.

Session Lifetime and Machine Specs

• Idle timeout: ~60 minutes (configurable somewhat via settings)
• Maximum session duration: 12 hours (then restarts)
• Machine type: usually e2-micro or similar (1 vCPU, ~3.75 GB RAM — can vary)
• Ephemeral disk: ~100 GB (non-persistent)

Quotas and Limits (as of late 2025 / early 2026)

• Persistent storage: 5 GB per user (hard limit)
• Weekly Cloud Code / Editor usage: 50 hours (soft quota for extended editor usage)
• No custom machine type selection
• No sudo access for installing system packages (apt install works for user-space tools)

Key Features and Tools Included

• Full gcloud CLI (authenticated automatically)
• Cloud Shell Editor — click pencil icon or type cloudshell edit <file>
• Git cloning, pushing (GitHub, Cloud Source Repositories)
• Docker (build & run containers — but no privileged mode)
• Terraform, Pulumi, Ansible, etc.
• Web preview: click port number → “Preview on port xxxx”
• File upload/download via UI (gear icon)
• Multiple tabs (in popped-out window)
• Customize prompt, install oh-my-zsh, etc.

Best Practices

1. Always store important data in persistent home or GCS

   gsutil cp -r important-project gs://my-backup-bucket/
   

2. Use .bashrc / .zshrc wisely

   Add aliases, functions, PATH modifications.

3. Install user tools in ~/bin or use language package managers

   pip install --user awscli
   npm install -g serverless
   go install github.com/...
   

4. Use Cloud Storage for larger files / shared data

   gcsfuse my-bucket ~/mnt/bucket
   

5. Leverage “Open in Cloud Shell” buttons in docs/tutorials

6. Use tmux / screen for long-running processes

   (session survives browser close if you re-attach)

7. Backup dotfiles

   Push ~/.* configs to private Git repo.

8. For heavy development → consider Cloud Workstations or local setup

   Cloud Shell is great for quick tasks, not full-time heavy dev.

Common Commands to Get Started

# Show current project
gcloud config get-value project

# List projects
gcloud projects list

# Set default project
gcloud config set project my-project-id

# List buckets
gsutil ls

# Clone repo and open editor
git clone https://github.com/user/repo
cloudshell edit repo/main.py

# Preview web app
python3 -m http.server 8080   # → click "Preview on port 8080"

When NOT to Use Cloud Shell

• Heavy compute (ML training, video encoding)
• Needing >5 GB persistent storage
• Requiring sudo / kernel modules
• Long-running background processes (>12h)
• Very large file handling

For those cases → use Compute Engine, Cloud Run, Cloud Workstations, or local machine + Cloud SDK.

Cloud Shell remains one of the fastest ways to interact with Google Cloud securely from any device with a browser.

References:

• Cloud Shell documentation
• Use Cloud Shell
• How Cloud Shell works
• Launch Cloud Shell
• Cloud Shell Editor limitations

---

Back

x-ai/grok-4.1-fast

Donate