# IPI Provisioner MCP Server

An MCP (Model Context Protocol) server for OpenShift Baremetal IPI (Installer-Provisioned Infrastructure) cluster provisioning. This tool helps you generate, validate, and manage install-config.yaml files, test BMC connectivity, validate network configurations, and guide you through the cluster installation process.

## Features

- **Install Config Management**: Generate and validate install-config.yaml files
- **BMC Testing**: Test Redfish/IPMI BMC connectivity
- **Network Validation**: Validate CIDR ranges, VIPs, and DNS configuration
- **Cluster Lifecycle**: Extract installer, create manifests, and manage installations
- **Templates**: Pre-built templates for Compact and HA deployments

## Architecture

```
ipi-provisioner-mcp/
├── mcp_server/
│   ├── config.py              # Configuration and environment variables
│   ├── main.py                # FastMCP server setup
│   ├── models/                # Pydantic data models
│   │   ├── install_config.py  # InstallConfig models
│   │   ├── baremetal.py       # Baremetal host models
│   │   └── network.py         # Network configuration models
│   ├── services/              # Business logic
│   │   ├── validation_service.py   # Config validation
│   │   ├── bmc_service.py          # BMC operations
│   │   ├── network_service.py      # Network validation
│   │   └── openshift_service.py    # Installer operations
│   ├── tools/                 # MCP tool implementations
│   │   ├── config_tools.py    # Config generation/validation
│   │   ├── bmc_tools.py       # BMC testing
│   │   ├── network_tools.py   # Network validation
│   │   └── cluster_tools.py   # Cluster lifecycle
│   └── utils/                 # Utilities
│       ├── yaml_handler.py    # YAML operations
│       └── templates.py       # Template management
├── templates/                 # Install-config templates
│   ├── baremetal-ipi-compact.yaml
│   └── baremetal-ipi-ha.yaml
└── tests/                     # Unit tests
```

## Available MCP Tools

### Config Generation & Validation (6 tools)

1. **`validate_install_config`** - Validate install-config.yaml content
2. **`generate_install_config`** - Generate install-config.yaml from parameters
3. **`get_config_template`** - Get a pre-built template
4. **`list_config_templates`** - List all available templates
5. **`create_baremetal_host_entry`** - Create a host entry for install-config

### BMC Operations (4 tools)

1. **`test_bmc_connection`** - Test single BMC connectivity (Redfish)
2. **`validate_bmc_addresses`** - Validate multiple BMC addresses
3. **`parse_bmc_address`** - Parse and validate BMC address format
4. **`generate_bmc_test_script`** - Generate bash script for BMC testing

### Network Validation (8 tools)

1. **`validate_network_cidr`** - Validate CIDR notation
2. **`check_ip_in_network`** - Check if IP is in CIDR range
3. **`check_network_overlap`** - Check if two networks overlap
4. **`validate_vip_configuration`** - Validate VIP configuration
5. **`generate_dns_config`** - Generate DNS records for cluster
6. **`calculate_cluster_network_capacity`** - Calculate pod network capacity
7. **`generate_network_diagram`** - Generate network diagram
8. **`validate_complete_network_config`** - Comprehensive network validation

### Cluster Lifecycle (6 tools)

1. **`check_installer_binary`** - Check if openshift-install exists
2. **`extract_openshift_installer`** - Extract installer from release image
3. **`create_installation_manifests`** - Create Kubernetes manifests
4. **`prepare_cluster_installation`** - Prepare installation command
5. **`prepare_cluster_destroy`** - Prepare cluster destroy command
6. **`get_installation_status`** - Get installation status
7. **`get_installation_logs`** - Get installation logs

## Installation

```bash
cd /Users/jhajyahy/mcp/ipi-provisioner-mcp
uv sync  # Creates venv and installs dependencies
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
```

## Configuration

### Environment Variables (Optional)

```bash
# OpenShift Configuration
export OPENSHIFT_RELEASE_IMAGE="quay.io/openshift-release-dev/ocp-release:4.17.0-x86_64"
export PULL_SECRET_PATH="$HOME/pull-secret.json"
export SSH_KEY_PATH="$HOME/.ssh/id_rsa.pub"

# MCP Transport (default: stdio)
export MCP_TRANSPORT="stdio"  # or "sse" for web-based integration
export MCP_HOST="127.0.0.1"   # for sse transport
export MCP_PORT="8000"         # for sse transport
```

### MCP Client Configuration

#### Claude Desktop / Cursor IDE

Add to your `~/.config/claude-desktop/mcp.json` or `~/.cursor/mcp.json`:

```json
{
  "mcpServers": {
    "ipi-provisioner": {
      "command": "uv",
      "args": ["run", "python", "/Users/jhajyahy/mcp/ipi-provisioner-mcp/main.py"],
      "description": "IPI Cluster Provisioning Assistant for OpenShift Baremetal",
      "env": {
        "OPENSHIFT_RELEASE_IMAGE": "quay.io/openshift-release-dev/ocp-release:4.17.0-x86_64",
        "PULL_SECRET_PATH": "/path/to/pull-secret.json",
        "SSH_KEY_PATH": "/path/to/id_rsa.pub"
      }
    }
  }
}
```

## Usage Examples

### 1. List Available Templates

```
User: List available install-config templates
```

Response shows Compact and HA templates with descriptions.

### 2. Generate Install Config

```
User: Generate an install-config for a 3-master HA cluster named 'prod-cluster'
with base domain 'example.com', API VIP 192.168.1.10, Ingress VIP 192.168.1.11
```

### 3. Validate Install Config

```
User: Validate this install-config.yaml:
[paste your config]
```

Returns validation errors, warnings, and a summary.

### 4. Test BMC Connectivity

```
User: Test BMC connection to redfish://192.168.1.100 with username admin and password secret
```

### 5. Validate Network Configuration

```
User: Validate network config with cluster CIDR 10.128.0.0/14, service CIDR 172.30.0.0/16,
machine CIDR 192.168.1.0/24, API VIP 192.168.1.10, Ingress VIP 192.168.1.11
```

### 6. Generate DNS Configuration

```
User: Generate DNS config for cluster 'ocp' with base domain 'example.com',
API VIP 192.168.1.10, Ingress VIP 192.168.1.11, and hosts:
- master-0 with IP 192.168.1.20
- master-1 with IP 192.168.1.21
- master-2 with IP 192.168.1.22
```

This will generate A and PTR records for all cluster components including the API VIP, Ingress VIP, and all hosts.

### 7. Extract OpenShift Installer

```
User: Extract openshift-install from release image
quay.io/openshift-release-dev/ocp-release:4.17.0-x86_64
```

### 8. Check Installation Status

```
User: Get installation status for /path/to/install-dir
```

## Common Workflows

### Workflow 1: Validate Existing Config

1. Validate config: `validate_install_config(yaml_content)`
2. Validate networks: `validate_complete_network_config(...)`
3. Validate BMCs: `validate_bmc_addresses(hosts)`
4. Generate DNS config: `generate_dns_config(...)`

### Workflow 2: Troubleshoot Installation

1. Check status: `get_installation_status(install_dir)`
2. Get logs: `get_installation_logs(install_dir, lines=100)`
3. Validate network: `validate_complete_network_config(...)`
4. Test BMCs: `validate_bmc_addresses(hosts)`

## Testing

Run the test suite:

```bash
uv run pytest tests/ -v
```

## Templates

### Available Templates

1. **baremetal-ipi-compact.yaml** - Compact Cluster
   - 3 master nodes (also act as workers)
   - Requires API and Ingress VIPs
   - Good for resource-constrained environments

2. **baremetal-ipi-ha.yaml** - High Availability
   - 3 master nodes
   - 2+ dedicated worker nodes
   - Requires API and Ingress VIPs
   - Production-ready configuration

## Network Requirements

### Machine Network
- Physical network for baremetal hosts
- Contains API VIP and Ingress VIP (for HA)
- Example: 192.168.1.0/24

### Cluster Network (Pod Network)
- Virtual network for pod-to-pod communication
- Example: 10.128.0.0/14
- Each node gets a /23 subnet (510 IPs per node)

### Service Network
- Virtual network for Kubernetes services
- Example: 172.30.0.0/16

### DNS Requirements (HA/Compact)

#### A Records
- `api.<cluster_name>.<base_domain>` → API VIP
- `*.apps.<cluster_name>.<base_domain>` → Ingress VIP (wildcard)
- `<hostname>.<base_domain>` → Host IP (for each cluster node)

#### PTR Records (Reverse DNS)
- API VIP → `api.<cluster_name>.<base_domain>`
- Each host IP → `<hostname>.<base_domain>`

Example for cluster 'ocp' with domain 'example.com':
```
# A Records
api.ocp.example.com.        A    10.1.235.200
*.apps.ocp.example.com.     A    10.1.235.201
master-0.example.com.       A    10.1.235.25
master-1.example.com.       A    10.1.235.26
master-2.example.com.       A    10.1.235.27

# PTR Records
200.235.1.10.in-addr.arpa.  PTR  api.ocp.example.com.
25.235.1.10.in-addr.arpa.   PTR  master-0.example.com.
26.235.1.10.in-addr.arpa.   PTR  master-1.example.com.
27.235.1.10.in-addr.arpa.   PTR  master-2.example.com.
```

## BMC Requirements

### Supported BMC Types
- Redfish (recommended)
- Redfish Virtual Media
- IPMI
- iDRAC Virtual Media

### BMC Address Format
- Redfish: `redfish://192.168.1.100` or `redfish+https://192.168.1.100:443`
- IPMI: `ipmi://192.168.1.100`

## Troubleshooting

### Common Issues

1. **BMC Connection Failures**
   - Use `test_bmc_connection` to verify connectivity
   - Check BMC credentials
   - Verify network connectivity to BMC

2. **Network Validation Errors**
   - Use `validate_complete_network_config` to check for overlaps
   - Ensure VIPs are in machine network range
   - Verify DNS is configured correctly

3. **Installation Failures**
   - Use `get_installation_logs` to check logs
   - Verify all BMCs are reachable
   - Check network configuration
   - Ensure DNS records exist (for HA)

## Contributing

Contributions are welcome! Please ensure:
- All tests pass: `uv run pytest tests/ -v`
- Code follows existing patterns
- Add tests for new functionality

## License

Apache 2.0

---

**Built for OpenShift Metal Platform Team**