When sACN universes will not arrive

The screenshot above is not a healthy Bridge. It carries an error, and it is the single most likely thing to go wrong on a laptop:

sACN: could not join 12 sACN universe(s) (2, 3, 4, 11, 12, 13, and 6 more);
they will not be received — bind a single network interface,
or raise net.ipv4.igmp_max_memberships

What is happening. sACN is multicast: one multicast group per universe. To receive a universe, the Bridge joins its group — and with the interface set to Automatic · all interfaces, it joins once per multicast-capable interface. A 13-universe scene on a machine with four such interfaces is 52 joins, plus the discovery universe.

Linux caps the number of multicast memberships one socket may hold at net.ipv4.igmp_max_memberships, 20 by default. Everything past the cap silently fails to join, and a universe with no membership is simply never delivered — which at the desk reads as a console that keeps dropping out rather than as a receiver that never subscribed. Hence the error: it is said once, where you can see it.

The interface count is what catches people out. A developer laptop happily carries eth0, wlan0, docker0, a handful of br-… bridges and a virbr0 — none of which will ever see a lighting console, and every one of which eats memberships.

Fix 1 — bind one interface (do this first)

Set Network interface to the interface that is actually on the lighting network, instead of Automatic · all interfaces. Joins drop to one per universe, the docker and VM bridges stop consuming the budget, and you also stop listening to networks the console is not on.

This is the right fix in almost every case, and it needs no root.

Fix 2 — raise the kernel limit

If you genuinely need many universes across more than one interface, raise the cap. Try it first:

sudo sysctl -w net.ipv4.igmp_max_memberships=256

and make it survive a reboot:

echo 'net.ipv4.igmp_max_memberships=256' | sudo tee /etc/sysctl.d/99-flexviz-sacn.conf
sudo sysctl --system

Restart the Bridge afterwards — memberships are joined when the receiver starts, so a running Bridge does not pick up the new limit.

Check the current value with:

sysctl net.ipv4.igmp_max_memberships

What this does not affect

• Art-Net is unaffected. It is broadcast and unicast, and joins no multicast groups.

• Unicast sACN still arrives regardless of membership. If your console sends unicast rather than multicast, this error does not apply to you.

• Windows and macOS have their own, generally higher, limits — this particular error is a Linux one.