Without multicast snooping working properly, all VMs connected to a network would receive all multicast traffic. On a big and busy network, this puts unnecessary stress on its members.

While with Linux bridge multicast snooping is enabled by default, that is not the case with OVS (used with localnet). In order to guide our users to good defaults, but to not change behaviour of already defined networks, we should explicitly enable this feature in our documentation examples.

4.16+ (older versions have a NetworkManager bug in snooping configuration)

Docs

1. Go to https://docs.redhat.com/en/documentation/openshift_container_platform/4.18/html-single/networking/index#configuration-localnet-switched-topology_configuring-additional-network-ovnk
2. See the second NodeNetworkConfigurationPolicy example

      type: ovs-bridge
      state: up
      bridge:
        allow-extra-patch-ports: true
        options:
          stp: false

      type: ovs-bridge
      state: up
      bridge:
        allow-extra-patch-ports: true
        options:
          stp: false
          mcast-snooping-enable: true <1>


<1> Multicast snooping prevents network devices from flooding multicast traffic to all members of the network. `false` by default

The same adjustement should be made to the examples including br-ex (type: ovs-bridge) in https://docs.redhat.com/en/documentation/openshift_container_platform/4.18/html/installing_on_bare_metal/user-provisioned-infrastructure#creating-manifest-file-customized-br-ex-bridge_installing-bare-metal