fix(mdns-relay): turn off RA on net1 so the IPv6 forward uses eth0 #1598

Merged
k.renner merged 1 commits from fix/mdns-relay-no-ra into main 2026-11-07 19:43:47 +00:00
2 changed files with 13 additions and 1 deletions
@@ -44,7 +44,7 @@ spec:
annotations:
# Bump to restart the pod and re-attach net1 (DHCP via multus).
# 2026-10-06: first attach on containernetworking/plugins v1.9.1.
homelab.renner.uno/restarted-for: "2026-10-06 cni-plugins v1.9.1 attach check"
homelab.renner.uno/restarted-for: "2026-11-07 net1 without RA (S28-36)"
k8s.v1.cni.cncf.io/networks: mdns-relay-iot
spec:
# Runs where the node patch declares the IoT VLAN (ens19).
@@ -31,6 +31,11 @@ spec:
# dropped. The relay's forward path to Home Assistant runs over eth0, so
# that route has to stay intact.
#
# sbr covers IPv4 only. With Router Advertisements net1 gets a global IPv6
# address and default route in the main table, and the IPv6 forward to the
# Service leaves via net1 with an IoT source, which HA's CNP drops. tuning
# turns RA processing off; the link-local address ff02::fb needs stays.
#
# Requires the cni-dhcp-daemon DaemonSet in infrastructure/multus.
config: |
{
@@ -45,6 +50,13 @@ spec:
"type": "dhcp"
}
},
{
"type": "tuning",
"sysctl": {
"net.ipv6.conf.IFNAME.accept_ra": "0",
"net.ipv6.conf.IFNAME.autoconf": "0"
}
},
{
"type": "sbr"
}