pfSense

In 23.01/2.7 the QinQ is handled by if_vlan directly and not netgraph. It now uses s-tags by default.

Since that's a change from the previous config it may fail for some setups making a setting there more important.

It is trivial to set at the command line though:

[23.01-DEVELOPMENT][admin@7100.stevew.lan]/root: ifconfig lagg0.4080 vlanproto 802.1Q
[23.01-DEVELOPMENT][admin@7100.stevew.lan]/root: ifconfig lagg0.4080
lagg0.4080: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
    options=4600003<RXCSUM,TXCSUM,RXCSUM_IPV6,TXCSUM_IPV6,NOMAP>
    ether 00:08:a2:0e:a5:93
    inet6 fe80::208:a2ff:fe0e:a593%lagg0.4080 prefixlen 64 scopeid 0x17
    groups: vlan
    vlan: 4080 vlanproto: 802.1q vlanpcp: 0 parent interface: lagg0
    media: Ethernet autoselect
    status: active
    nd6 options=21<PERFORMNUD,AUTO_LINKLOCAL>

Steve Wheeler wrote in #note-1:

In 23.01/2.7 the QinQ is handled by if_vlan directly and not netgraph. It now uses s-tags by default.

Since that's a change from the previous config it may fail for some setups making a setting there more important.

It is trivial to set at the command line though:
[...]

We faced this issue in our recent 2.6 -> 2.7 upgrade and thank you for posting this! Saved our butts. Whats the most elegant way to make this persistent across reboots and even better yet done in UI ?

We faced this issue in our recent 2.6 -> 2.7 upgrade and thank you for posting this! Saved our butts. Whats the most elegant way to make this persistent across reboots and even better yet done in UI ?

Well, this issue just bit me too. My provider doesn't support ethertype 0x88A8 on their QinQ service, so upgrading broke all of my QinQ connections.

The command line fix worked, for now. But now every reboot will break everything.

Installing the shellcmd package and adding that command to run on boot makes things work. Not an elegant solution but at least we're back up and running.