Fix unreachable IPv6 RHEL-based Instances after transitioning to routed IP

Reviewed on 21 February 2024 • Published on 16 November 2023

RHEL-based Instances created before November 15 2023 require a NetworkManager configuration that enables correct IPv6 configuration, after a transition to routed IP is performed. If your Instance is configured only with an IPv6 address, it will become unreachable after the migration. It is important to add the NetworkManager configuration so IPv6 configuration remains correct.

The affected O/S distributions are:

CentOS Stream versions 8 and 9
RockyLinux versions 8 and 9
AlmaLinux versions 8 and 9
Fedora versions 37 and 38

Note

CentOS 7.9 is not concerned by this issue as it does not support routed IP.

Your Instance has not been migrated to routed IP yet

Connect to your Instance and execute the following command as root user to add the required NetworkManager option:

cat << EOF > /etc/NetworkManager/conf.d/00-scaleway.conf
[connection]
# The value 0 stands for eui64 -- see nm-settings-nmcli(5)
ipv6.addr-gen-mode=0
EOF

Your Instance has been moved to routed IP

If your Instance has been moved to routed IP and you have IPv4 connectivity, add the NetworkManager configuration as outlined in the previous section. After the next reboot, you will recover IPv6 connectivity while IPv4 configuration will remain untouched

You only have IPv6 configured so you cannot connect to the Instance anymore

Important

Two reboots of your Instance will be required in order to recover IPv6 connectivity.

If you are in this situation, you need to temporarily add an IPv4 address to connect to your Instance in order to add the NetworkManager configuration.

scw instance server list
ID            NAME    TYPE    STATE     ZONE      PUBLIC IP           PRIVATE IP  TAGS  IMAGE NAME          ROUTED
{Instance_ID} rocky8  DEV1-S  running   fr-par-2  2001:bc8:1210:26a:dc00:ff:fe23:ab83  - [] RockyLinux 8     true
scw instance ip create type=routed_ipv4 server={Instance_ID}
scw instance server reboot {Instance_ID}

Once the reboot is completed, list your IP addresses and identify the IPv4 address attached to your Instance:

$ scw instance ip list
ID                 ADDRESS         PREFIX                  TYPE         REVERSE  PROJECT       SERVER ID     SERVER NAME  TAGS  ZONE
{IPv6_address_ID}  -               2001:bc8:1210:26a::/64  routed_ipv6  -        {project id}  {project id}  rocky8       []    fr-par-2
{IPv4_address_ID}  51.159.143.128  -                       routed_ipv4  -        {project id}  {project id}  rocky8       []    fr-par-2

You can now connect to your Instance with the IPv4 address to make the NetworkManager configuration change:

ssh root@51.159.143.128
cat << EOF > /etc/NetworkManager/conf.d/00-scaleway.conf
[connection]
ipv6.addr-gen-mode=0
EOF

Once modified, reset your Instance to its prior configuration. It is necessary to first detach the IPv4 address before attaching the IPv6 address.

scw instance server standby {Instance_ID}
scw instance server update {Instance_ID} ip=none
scw instance server update {Instance_ID} ip={IPv6_address_ID}
scw instance server start {Instance_ID}
scw instance ip delete {IPv4_address_ID}

You can now connect using your IPv6 address:

ssh root@2001:bc8:1210:26a:dc00:ff:fe23:ab83
Warning: Permanently added '2001:bc8:1210:26a:dc00:ff:fe23:ab83' (ED25519) to the list of known hosts.
Activate the web console with: systemctl enable --now cockpit.socket
 Last login: Wed Nov 15 16:53:09 2023 from 51.159.46.153
 [root@rocky8 ~]#