Skip to content

LXD Infrastructure & Network Configuration

This document specifies the actual infrastructure state and networking configuration of the LXD Hypervisor Cluster backing the Intelligence Cloud Platform, as retrieved directly from the live LXD daemon API.

1. Hypervisor Cluster Architecture

The platform operates on a clustered hypervisor stack managed via LXD and MicroCloud.

  • Cluster Daemon: LXD v5.21.4 (LTS)
  • Clustering Engine: MicroCloud v2.1.3
  • Clustering Status: Clustered (true)
  • Firewall Driver: nftables
  • Storage Pools: ZFS (local-pool backed by zfs-2.2.2)
  • OVN Integration: Open Virtual Network (OVN) overlay network backed by physical interfaces and managed database controllers.
    • OVN Northbound Connections:
      • ssl:10.10.10.13:6641
      • ssl:10.10.10.14:6641
      • ssl:10.10.10.10:6641

[!NOTE] OVN Northbound Connections refer to the endpoints through which the LXD daemon communicates with the OVN Northbound Database (NB DB) over secured SSL ports (6641). The NB DB stores high-level logical network structures (switches, routers, load balancers, NAT). In this cluster, nodes c03 (10.10.10.13), c04 (10.10.10.14), and cloud (10.10.10.10) maintain a replicated database cluster using Raft consensus, ensuring active-standby database failover and high-availability network provisioning.

Cluster Nodes (Members) & Exit Nodes

The hypervisor cluster consists of the following 7 registered member nodes:

  • c03 (IP: 10.10.10.13) — Exit Node (ovn-chassis), Database Standby.
  • c04 (IP: 10.10.10.14) — Exit Node (ovn-chassis), Database Member.
  • c05 (IP: 10.10.10.15) — Exit Node (ovn-chassis), Database Leader.
  • c06 (IP: 10.10.10.16) — Exit Node (ovn-chassis), Database Standby.
  • cloud (IP: 10.10.10.10) — Primary Entry Node, Database Member.
  • c20 (IP: 10.10.10.20) — Worker Node (secondary group).
  • c22 (IP: 10.10.10.22) — Worker Node (secondary group).

[!NOTE] The nodes c03, c04, c05, and c06 are assigned the ovn-chassis role in the cluster. These four nodes act as the physical OVN gateway exit nodes to the internet for all OVN virtual networks (such as ovn-net) mapped onto the UPLINK network.

2. Hypervisor Network Interfaces

The following table summarizes the managed virtual and physical networks configured inside the hypervisor:

Network Name Type Subnet / Configuration NAT Purpose
UPLINK physical Gateway: 147.27.50.254/24
Nameservers: 147.27.18.1, 8.8.8.8		 false Physical connection bridge to external networks.
Hosts static route allocations and external OVN IP mapping spaces.
ovn-net ovn Subnet: 10.0.50.1/24
Volatile IP: 147.27.50.30		 true Standard tenant virtual overlay network. Isolated Layer-2 tenant domains.
transit-net bridge Subnet: 10.254.0.1/24
DHCP Range: 10.254.0.10-10.254.0.20
OVN Range: 10.254.0.130-10.254.0.250		 true Layer-3 transit bridge routing traffic between OVN segments and standard bridge networks.
lxdfan0 bridge Underlay: 147.27.50.0/24 true Local Fan bridge networking across nodes in default development configurations.