The bug is about multiple points:

1. https://docs.openshift.com/container-platform/4.17/networking/metallb/metallb-configure-return-traffic.html#nw-metallb-configure-return-traffic-proc_metallb-configure-return-traffic

in the NNCP provided (below the "with content like the following example:" part), we need to add another route-rule at the bottom (depending on the version:
* FOR 4.16:
      - ip-to: 169.254.169.0/29
        priority: 998
        route-table: 254

* FOR 4.17+:
      - ip-to: 169.254.0.0/17
        priority: 998
        route-table: 254

In addition, we should change point no. 9 to include the new rule:
The ip-to fields must match the Cluster Network CIDR, Service Network CIDR and Internal Masquerade Subnet CIDR. You can view the values for these CIDR address specifications by running the following command: oc describe network.operator/cluster.

(we add a reference to Internal Masquerade Subnet AND changing the oc describe to point to network.operator/cluster)



2. https://docs.openshift.com/container-platform/4.17/networking/metallb/metallb-configure-return-traffic.html#nw-metallb-configure-return-traffic-proc_metallb-configure-return-traffic

in the EgressService provided ("Create an EgressService CR" part), point no 5. should be "Specify the routing table id for egress traffic.", that is make it clear it is the ID that should be specified. preferably, it should be clear that it should have the same value as in "route-table-id: 2" from the NNCP.



3. https://docs.openshift.com/container-platform/4.17/networking/ovn_kubernetes_network_provider/configuring-egress-traffic-for-vrf-loadbalancer-services.html#nw-egress-service-ovn-cr_configuring-egress-traffic-loadbalancer-services

similar to the previous point, under point no .5 we add the id: "Optional: Specify the routing table id for egress traffic. ..." 
(we make it clear that it is the ID that should be specified, here we don't have an NNCP to reference).

    1.
    2.
    3.