Skip to content

Network ‐ OVN Routing Health Check

hoberger-rh edited this page Apr 15, 2026 · 3 revisions

Description

This rule checks that the routing table contains routes via OVN management interfaces (ovn-k8s-mp0, ovn-k8s-mp1, etc.), which are essential for:

  • Pod-to-pod communication
  • Pod-to-service communication
  • Cluster network connectivity
  • Node-to-API server connectivity

Prerequisites

  • OpenShift cluster with OVN-Kubernetes networking
  • OVN controller running on node
  • OVN management interface created (ovn-k8s-mp)
  • Command: ip

Impact

Missing OVN routes cause complete network failure:

  • Pod communication failure - Pods cannot communicate with other pods
  • Service networking broken - Pods cannot reach Kubernetes services
  • Node NotReady state - Node cannot reach API server
  • DNS resolution failure - Pods cannot reach DNS (CoreDNS)
  • Application failures - All pod workloads fail networking
  • Cluster degradation - Node effectively isolated from cluster

Root Cause

Common scenarios causing missing OVN routes:

  • OVN Controller Not Running
  • OVN Management Interface Missing
  • OVN Database Connectivity Issues
  • OVN Configuration Errors
  • System Routing Table Corruption

Diagnostics

Check the routing table for routes via OVN management interface:

# Display all routes
ip route show

# Look for routes via ovn-k8s-mp0 or ovn-k8s-mp1
ip route show | grep ovn-k8s-mp

# Check if OVN management interface exists and is UP
ip link show | grep ovn-k8s-mp
ip link show ovn-k8s-mp0

The node should have routes via the OVN management interface (ovn-k8s-mp0 or ovn-k8s-mp1).

Resources

Clone this wiki locally