.In IPv6 environments, Networking service erroneously configures default route in response to router advertisements Currently, in IPv6 environments, the IPv6 key-value pair, `IPV6_AUTOCONF=no`, in `ifcfg-*` files generated by `os-net-config` does not prevent the system from configuring the default route in response to router advertisements. There are two approaches to the workaround. In Approach 1, you edit the configuration files directly and then run `sysctl`. In Approach 2, you run the `sysctl` command to activate and write these values to the `sysctl.conf` file. *Workaround Approach 1:* Edit the configuration files . Set `net.ipv6.conf.<interface>.accept_ra_defrtr=0` to prevent learning the default route from router advertisements and `net.ipv6.conf.<interface>.accept_ra=0` to prevent the system from accepting router advertisements for any routes. + Set both of these variables in `/etc/sysctl.conf` or `/etc/sysctl.d/99-sysctl.conf` (or another file in that subdirectory). + For each specific interface: + ---- net.ipv6.conf.<interface>.accept_ra_defrtr=0 net.ipv6.conf.<interface>.accept_ra=0 ---- + [NOTE] ==== Setting each specific interface can help to ensure that the settings are not overridden. For newly created interfaces: ---- net.ipv6.conf.default.accept_ra_defrtr=0 net.ipv6.conf.default.accept_ra=0 ---- For all interfaces present at boot time: ---- net.ipv6.conf.all.accept_ra_defrtr=0 net.ipv6.conf.all.accept_ra=0 ---- ==== . Run the following command: + ---- $ sudo sysctl -p ---- *Workaround Approach 2:* run `sysctl` to activate and write the values to the `sysctl.conf` file: * For each specific interface: + ---- $ sudo sysctl -w net.ipv6.conf.<interface>.accept_ra_defrtr=0 $ sudo sysctl -w net.ipv6.conf.<interface>.accept_ra=0 ---- + [NOTE] ==== Setting each specific interface can help to ensure that the settings are not overridden. For newly created interfaces: ---- $ sudo sysctl -w net.ipv6.conf.default.accept_ra_defrtr=0 $ sudo sysctl -w net.ipv6.conf.default.accept_ra=0 ---- For all interfaces present at boot time: ---- $ sudo sysctl -w net.ipv6.conf.all.accept_ra_defrtr=0 $ sudo sysctl -w net.ipv6.conf.all.accept_ra=0 ---- ====

This behavior occurs only in RHEL9 and not RHEL8, and only when using ifcfg but not when using nmstate. A workaround has been documented using sysctl settings. The precise kernel changes that changed the behavior has been requested from the kernel team (issue linked), but an answer may take some time.