# QEMU Socket Networking VM Cluster

## Architecture

**Topology:** 4 QEMU VMs connected via multicast socket networking (230.0.0.1:1234)

**VMs:**
1. **pxe-server** (192.168.100.1) - Provides DHCP/TFTP/HTTP services
2. **node01** (192.168.100.11) - Cluster node
3. **node02** (192.168.100.12) - Cluster node
4. **node03** (192.168.100.13) - Cluster node

**Network:** All VMs share L2 segment via QEMU multicast socket (no root privileges required)

## Files

- `node01.qcow2`, `node02.qcow2`, `node03.qcow2` - 100GB cluster node disks
- `pxe-server.qcow2` - 20GB PXE server disk
- `launch-pxe-server.sh` - PXE server startup script
- `launch-node01.sh`, `launch-node02.sh`, `launch-node03.sh` - Node startup scripts
- `pxe-server/` - PXE server configuration files

## MACs

- pxe-server: 52:54:00:00:00:01
- node01: 52:54:00:00:01:01
- node02: 52:54:00:00:01:02
- node03: 52:54:00:00:01:03

## Provisioning Flow

1. Start PXE server VM (Alpine Linux with dnsmasq)
2. Configure DHCP/TFTP/HTTP services
3. Deploy NixOS netboot artifacts
4. Start node VMs with PXE boot enabled
5. Nodes PXE boot and provision via nixos-anywhere