iptables-restore fails with error '"Failed to execute iptables-restore" err="exit status 2 (iptables-restore v1.8.4 (nf_tables): Port `:ffff:x.x.x.x]:yyyy' not valid\n\nError occurred at line: 645"'

Issue

• If a user manually crafts a Endpoint object with field subsets.addresses.ip containing a IPv4-mapped IPv6 address (example: ::ffff:10.128.0.10) and then recreates a SDN pod, the SDN pod will fail to run iptables-restore showing errors such as the following (note that the IP is ::ffff:10.128.0.10 while the error shows an address with a truncated first colon :ffff:10.128.0.10):

E0303 16:14:46.953649 140364 proxier.go:1567] "Failed to execute iptables-restore" err="exit status 2 (iptables-restore v1.8.4 (nf_tables): Port `:ffff:10.128.0.10]:8080' not valid\n\nError occurred at line: 645\nTry `iptables-restore -h' or 'iptables-restore --help' for more information.\n

• After creating a Endpoint object with field subsets.addresses.ip containing a IPv4-mapped IPv6 address, accessing any Service in the cluster results in the error No route to host