Questions to Determine

• Is the Calico node pod running on the affected node?
• Are the CNI configuration files present in /etc/cni/net.d/?
• Are the CNI binary plugins present in /opt/cni/bin/?
• What do the kubelet logs show about CNI initialization?
• Is the VXLAN tunnel interface (vxlan.calico) present on the node?
• Did the kernel upgrade change the available kernel modules (vxlan, ip_tables)?
• Can the node reach other nodes on the VXLAN port (UDP 4789)?
• Is the calico-node DaemonSet pod healthy on this node?
• Are IP routes for pod CIDRs of other nodes present in the routing table?