Symptoms: Network Bonding Failover Not Working

• A production web server has a Linux bond (bond0) configured with two interfaces: eno1 and eno2.
• The bond is intended to provide active-backup failover for high availability.
• During a planned switch maintenance, the port for eno1 was shut down on the switch.
• Expected behavior: traffic fails over to eno2 seamlessly. Actual: the server loses all network connectivity.
• Pinging the server from the monitoring system shows 100% packet loss during the test.
• When the switch port for eno1 is re-enabled, connectivity returns within seconds.
• The bond was configured 6 months ago and this is the first time failover was actually tested.
• Both interfaces show "UP" in ip link output even when one switch port is shut.
• The server runs RHEL 8 and uses NetworkManager for network configuration.