KamoCRM

ICE on k3m1's LAN address only, so SIP legs stop resuming every 15 s

FixKlusterServices
Shipped
October 28, 2026 at 5:25 AM UTC
Author
Kamo
Commit
e9b8ba4

livekit-sip's same-host WebRTC legs to livekit-server failed ICE 15 s after connecting (the server's 10 s disconnected + 5 s failed timeouts) and resumed in a loop on 302+ legs from 02:04Z: the server offered every host address (eno1, dead wg0, docker0, cilium_host, the 10.0.50.1 link, the eno49 IPv6 ULA), so pairs landed on whichever won, the IPv6 ULA among them, whose responses came back late and were discarded. Pod clients already used 192.168.4.36; node_ip 47.181.8.87 is still advertised and TURN stays the fallback.

All changes

Like what you see shipping?

All of it arrives in your workspace on its own. Start on the free plan and read this page again in a month.

Start Free ForeverView Pricing