Skip to content

QUICKSTART.md

Latest commit

 

History

History
168 lines (112 loc) · 3.72 KB

QUICKSTART.md

File metadata and controls

168 lines (112 loc) · 3.72 KB

Quickstart — LISA MCP Server

Get from zero to AI-powered LISA testing in 10 minutes.

Prerequisites

  • Python 3.10 or later
  • Git
  • VS Code with GitHub Copilot (or any MCP-compatible client)

Step 1 — Clone both repositories

# Clone the LISA framework (the test suite library)
git clone https://github.com/microsoft/lisa.git ~/lisa

# Clone this MCP server
git clone https://github.com/kkkashan/LISA_MCP_Server.git ~/lisa-mcp-server

Step 2 — Install the MCP server

cd ~/lisa-mcp-server
pip install -e .

Verify the install:

python3 -c "from lisa_mcp.server import mcp; print('OK —', mcp.name)"
# OK — lisa-mcp-server

Step 3 — Register with VS Code

The repo ships with .vscode/mcp.json pre-configured. Open the workspace in VS Code and it will detect the MCP server automatically.

If you need to set the path manually, edit .vscode/mcp.json:

{
 "servers": {
 "lisa": {
 "command": "python3",
 "args": ["-m", "lisa_mcp.server"],
 "cwd": "/absolute/path/to/lisa-mcp-server"
 }
 }
}

Get your absolute path with:

cd ~/lisa-mcp-server && pwd

Step 4 — Start the server

In VS Code, open the Command Palette → MCP: Start Server → select lisa.

Or click the Start button that appears above the server entry in .vscode/mcp.json.

Step 5 — Your first commands

Open GitHub Copilot Chat (or your MCP client) and try:

5a. Check what test areas exist

Show me all the functional areas in my LISA repo at ~/lisa

The AI calls list_test_areas(lisa_path="~/lisa") and returns:

network, storage, cpu, memory, nvme, core, provisioning, hyperv, ...

5b. Discover tests by tier

Show me all priority-0 (T0) tests in the network area of ~/lisa

The AI calls discover_test_cases(lisa_path="~/lisa", area="network", tier="T0").

5c. Build a runbook

Build a T1 Azure runbook for Ubuntu 22.04 LTS and save it to ~/my_runbook.yml

The AI calls build_tier_runbook_file(tier="T1", platform_type="azure", output_path="~/my_runbook.yml").

5d. Generate a new test suite

Write a LISA test suite called "KernelSmokeTest" in the "cpu" area
that checks the kernel version is at least 5.15. Priority 0, Azure only.

The AI calls generate_test_suite_code(...) and returns complete Python source.

Step 6 — Run tests (requires LISA + Azure credentials)

Install LISA itself:

cd ~/lisa
pip install -e .
lisa --version

Then ask:

Run the runbook at ~/my_runbook.yml using the LISA repo at ~/lisa.
Pass subscription_id:xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
and admin_private_key_file:~/.ssh/id_rsa

The AI calls run_lisa_tests(...) and returns stdout/stderr + exit code.

Step 7 — Analyze failures with Azure OpenAI (optional)

If tests fail, get AI-powered root cause analysis:

Analyze the failures in ~/lisa/lisa_results.xml
Azure OpenAI key: <your-key>
Save the report to ~/reports/

The AI calls generate_analysis_report(...) and produces an HTML + Markdown report.

Azure OpenAI endpoint (pre-configured): https://kkopenailearn.openai.azure.com/openai/responses?api-version=2025-04-01-preview Pass your API key when calling any analyze_* tool.

What's next?