fix(bird): route LoadBalancer VIP replies into table 47 (#1547 )

A reply from a LoadBalancer VIP reaches the kernel's routing decision with the VIP as its source: Cilium reverses the service NAT in from-container before the packet enters the stack. The pod-source rules 1000/1001 do not match it, so it took main's default over mgmt0 while the request had arrived on ingress0 as the BGP next hop. bird's DaemonSet gets a vip-rules container that reads every enabled CiliumLoadBalancerIPPool and keeps one rule per block, priority 1002, `from <block> lookup 47`, reconciled every minute. The pool stays the only place the prefixes are written. Like the pod rules, a rule into an empty table 47 (bird gone) falls through to main. bgp-pool's blocks move from start/stop to the equivalent CIDRs so each block maps to one rule. Cilium v1.20.4 LB-IPAM compares a modified pool's blocks by first and last address (handlePoolModified, LBRange.EqualCIDR), and allowFirstLastIPs unset keeps .0/.255 and fd99::/fd99::ff allocatable, so the ranges are kept and no service is reallocated. Closes #1547
2026-11-02 06:36:08 +00:00
57 changed files with 196 additions and 158 deletions
@@ -47,7 +47,7 @@ when:

 steps:
  - name: build-ansible
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/ansible
    # Only rebuild when THIS image's source (or the pipeline file) changed, or
    # on a manual run. Without this every step ran on any matching push, so a
@@ -74,7 +74,7 @@ steps:
          --destination gitea.renner.uno/gitea_admin/ci-ansible:latest

  - name: build-k8s
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/k8s
    when:
      - event: deployment
@@ -97,7 +97,7 @@ steps:
          --destination gitea.renner.uno/gitea_admin/ci-k8s:latest

  - name: build-vmtest
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/vmtest
    when:
      - event: deployment
@@ -120,7 +120,7 @@ steps:
          --destination gitea.renner.uno/gitea_admin/ci-vmtest:latest

  - name: build-tofu
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/tofu
    when:
      - event: deployment
@@ -150,7 +150,7 @@ steps:
  # Runtime image, not CI tooling: a static musl binary in a FROM-scratch image,
  # used as the base for secret-rotation CronJobs. Replaces per-job inline Python.
  - name: build-secret-rotator
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: secret-rotator
    when:
      - event: deployment
@@ -175,7 +175,7 @@ steps:
  # review-bot: the Go service that reviews pull requests with a Mistral agent (webhook plus sweep).
  # Runs after Gitea is up, so no ghcr mirror.
  - name: build-review-bot
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: review-bot
    when:
      - event: deployment
@@ -225,7 +225,7 @@ steps:
  #
  # The gitea_admin/ copy stays until nothing references it any more.
  - name: build-alpine-jq
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/alpine-jq
    when:
      - event: deployment
@@ -276,7 +276,7 @@ steps:
  # note on build-alpine-jq. Anything an app needs while the cluster is still
  # being built cannot live only in that cluster's own registry.
  - name: build-cni-plugins
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/cni-plugins
    when:
      - event: deployment
@@ -316,7 +316,7 @@ steps:
  # gitea.renner.uno is not there to serve it. A cold-rebuild rehearsal left
  # paperless in Init:ErrImagePull on exactly this image.
  - name: build-alpine-openssl
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/alpine-openssl
    when:
      - event: deployment
@@ -351,7 +351,7 @@ steps:
  # Ansible converge. Not a ci-* image -- it is a runtime artifact -- but it
  # builds the same way and lands in the same registry so the audit sees it.
  - name: build-heating-can
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/heating-can
    when:
      - event: deployment
@@ -377,7 +377,7 @@ steps:
  # services (public Ingress hosts -> TLS passthrough -> Traefik ClusterIP).
  # Runtime artifact like heating-can; same registry so the audit tracks it.
  - name: build-netbird-service-sync
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/netbird-service-sync
    when:
      - event: deployment
@@ -402,7 +402,7 @@ steps:
  # authentik-gate-sync: the authentik-gate ingress class (Ingress -> Authentik
  # proxy gate + Traefik route). Its unit tests run as a stage of the image build.
  - name: build-authentik-gate-sync
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/authentik-gate-sync
    when:
      - event: deployment
@@ -428,7 +428,7 @@ steps:
  # kuma-sync: reconciles KumaMonitor resources into Uptime Kuma. Its unit
  # tests run as a stage of the image build.
  - name: build-kuma-sync
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/kuma-sync
    when:
      - event: deployment
@@ -454,7 +454,7 @@ steps:
  # cloudflare-lb-sync: reconciles CloudflareLoadBalancer resources into
  # Cloudflare. Its unit tests run as a stage of the image build.
  - name: build-cloudflare-lb-sync
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/cloudflare-lb-sync
    when:
      - event: deployment
@@ -480,7 +480,7 @@ steps:
  # bird: BGP daemon for the pod-network mesh over pod0 (#576). Runtime
  # artifact like heating-can; same registry so the audit tracks it.
  - name: build-bird
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/bird
    when:
      - event: deployment
@@ -504,7 +504,7 @@ steps:

  # Orchestrator for the monthly backup-verify CronJob (velero ns).
  - name: build-backup-verify
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/backup-verify
    when:
      - event: deployment
@@ -131,7 +131,7 @@ steps:
  #    --build-arg to the resolved version (not "latest") guarantees the tag and
  #    the image content agree even if a new release lands mid-run.
  - name: build
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/claude
    depends_on: [decide]
    environment:
@@ -172,7 +172,7 @@ steps:
  # whenever it trails the deployed server. If you bump the server, bump here.
  - name: woodpecker-lint
    depends_on: [] # see note above `steps:`
-    image: woodpeckerci/woodpecker-cli:v3.20.1-alpine@sha256:74c576faf49cce699825d11c7d7bd38ab5f007dd4f57cdbdd55099d2c0f2621a
+    image: woodpeckerci/woodpecker-cli:v3.19.0-alpine@sha256:27a77117e38d7bbdc380f0e90ff3ee4871ab1eefa78b5d9136a2c969fbecd9f4
    commands:
      - |
        set -eu
@@ -63,7 +63,7 @@ steps:
  # bridge, and the secret guard.
  - name: hermes-skill-tests
    depends_on: []
-    image: python:3.15-alpine@sha256:20a82dbc2783112ddb91d1534603f802d222a3e63e593cd002eb2ac534e70211
+    image: python:3.15-alpine@sha256:d44b1254cb92329418d164b22b20d86bbd3869d48e80e09e476531a409a728bb
    environment:
      PYTHONDONTWRITEBYTECODE: "1"
    commands:
@@ -10,7 +10,7 @@ when:

 steps:
  - name: build
-    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.20-debug@sha256:28af5a3bd9c51fedb353ad91de735eda53592b5c06a34e7f5a9550334725d7f7
+    image: registry.gitlab.com/gitlab-ci-utils/container-images/kaniko:v1.25.19-debug@sha256:7795cec509a50a71378cc2c365f55c555ee3a05834d8eb1cd2552df5e66d44d8
    directory: ci/images/hermes
    environment:
      # registry-bot write:package PAT (konairius org secret).
@@ -6,7 +6,7 @@ when:

 steps:
  - name: go-test
-    image: golang:1.27.3-bookworm@sha256:d9b6a1beb9474adc2a1ce842573639b44b2a0dd49b6669f51d3821baf9ec37c9
+    image: golang:1.27.3-bookworm@sha256:972b4110ba0f47539a975c31e52c64765e1ff19f2d74469ae4a595c6676e80a5
    commands:
      - test -z "$(gofmt -l review-bot)"
      - go -C review-bot vet ./...
@@ -5,12 +5,12 @@

 # The unit tests gate the image: the final stage copies the controller out of
 # this stage, so a failing test fails the build.
-FROM python:3.15-slim@sha256:94d3426f1def7db4a106a232ac121bfd975393e5dfd40a619b955fae45fe7913 AS test
+FROM python:3.15-slim@sha256:533b0b2f9b5d46fd097091cf826f27902f1d6c3bca714b6dc74551356aa97a51 AS test
 WORKDIR /src
 COPY controller.py test_controller.py ./
 RUN python3 -B -m unittest -v

-FROM python:3.15-slim@sha256:94d3426f1def7db4a106a232ac121bfd975393e5dfd40a619b955fae45fe7913
+FROM python:3.15-slim@sha256:533b0b2f9b5d46fd097091cf826f27902f1d6c3bca714b6dc74551356aa97a51

 RUN useradd --system --no-create-home --shell /usr/sbin/nologin controller
 COPY --from=test /src/controller.py /opt/controller.py
@@ -1,4 +1,4 @@
-FROM node:24-alpine@sha256:8a32c8cf649a49915a66e81e7a609223c708d0adabf288ee3f4077d765542f85
+FROM node:24-alpine@sha256:06d7adf3e0ef61a13f1d28682726366f97a15aa1888c934d3ecec86c33918b53

 # Default to the latest Claude Code. Pinning a fixed version is the wrong
 # default here: old releases eventually stop working against the API, so a
@@ -2,7 +2,7 @@
 # Cloudflare. The unit tests (no tags, so no envtest) gate the image: the
 # final stage copies the binary out of this stage, so a failing test fails the
 # build.
-FROM golang:1.27@sha256:2920bb124f1a2e12a1d39f0cab756ae0068441ee1da7655dec7f5e374944095b AS build
+FROM golang:1.27@sha256:f20d83a83e18dbd2a4714b8ffd4b30b287b6243962dd75608e86abcc63ba4efc AS build
 WORKDIR /src
 COPY go.mod go.sum ./
 RUN go mod download
@@ -10,7 +10,7 @@ COPY . .
 RUN go test ./...
 RUN CGO_ENABLED=0 go build -trimpath -ldflags="-s -w" -o /out/cloudflare-lb-sync ./cmd/cloudflare-lb-sync

-FROM gcr.io/distroless/static:nonroot@sha256:25ac8f0c3b2334a9eb05452266bd4697864a8c8ef0ef71e5e6480d7f92868329
+FROM gcr.io/distroless/static:nonroot@sha256:18f4cc26c9ffa7f20a359298be7deadac5c607c8e03a02b8119f1a5fc2f82c99
 COPY --from=build /out/cloudflare-lb-sync /cloudflare-lb-sync
 USER 65532
 ENTRYPOINT ["/cloudflare-lb-sync"]
@@ -23,7 +23,7 @@
 # namespace, so can0 is only reachable with network_mode: host. That is also
 # how the local MQTT broker on 127.0.0.1:1883 is reached. CAN_RAW sockets need
 # no capabilities, so both run unprivileged.
-FROM python:3.15-slim@sha256:94d3426f1def7db4a106a232ac121bfd975393e5dfd40a619b955fae45fe7913 AS build
+FROM python:3.15-slim@sha256:533b0b2f9b5d46fd097091cf826f27902f1d6c3bca714b6dc74551356aa97a51 AS build

 # Pins live here, in the artifact, rather than in Ansible variables.
 #
@@ -56,7 +56,7 @@ RUN git clone --quiet https://github.com/MyHomeMyData/E3onCAN.git /opt/e3oncan \
 && rm -rf /opt/e3oncan/.git


-FROM python:3.15-slim@sha256:94d3426f1def7db4a106a232ac121bfd975393e5dfd40a619b955fae45fe7913
+FROM python:3.15-slim@sha256:533b0b2f9b5d46fd097091cf826f27902f1d6c3bca714b6dc74551356aa97a51

 ARG OPEN3E_REF=6553c10453a6bae1210908f38a5af2481a46a3e5
 ARG E3ONCAN_REF=461dcbdf97e8d72d2edb9929c63b39382af6c3a1
@@ -6,10 +6,10 @@ RUN apt-get update \
 && apt-get install -y --no-install-recommends jq \
 && rm -rf /var/lib/apt/lists/*

-COPY --from=ghcr.io/siderolabs/talosctl:v1.14.5@sha256:0b852e35c1f513ee5eacc22183d6193486012420569fde39fcd5260c51938443 /talosctl /usr/local/bin/talosctl
-COPY --from=alpine/kubectl:1.37.2@sha256:5e2de75d37e811ba6eb92e06890b1b3263eae4f0a1c17b49269edec589b1352a /usr/local/bin/kubectl /usr/local/bin/kubectl
-COPY --from=ghcr.io/fluxcd/flux-cli:v2.10.0@sha256:1a5178b1770052d6e88ab1c7c8434c5ac9bca8a8ab984dcaf6b9df31814ce9f2 /usr/local/bin/flux /usr/local/bin/flux
-COPY --from=quay.io/openbao/openbao:2.7.2@sha256:5046911075e137e154ce49ab051502f4ec7cf367d484ae4a4dce665d9b90561f /usr/bin/bao /usr/local/bin/bao
+COPY --from=ghcr.io/siderolabs/talosctl:v1.14.4@sha256:b3233234327350f20ed194f3fad83d90a64344c1940f839c7b0c0abef50c8c33 /talosctl /usr/local/bin/talosctl
+COPY --from=alpine/kubectl:1.37.1@sha256:a5d0fbecab3f7b6c2ab7398f1daa73a3e0ed6ece18556d21698233e9e3759cc7 /usr/local/bin/kubectl /usr/local/bin/kubectl
+COPY --from=ghcr.io/fluxcd/flux-cli:v2.9.6@sha256:b1ac18156f227af9a524b842a96f2c20986c7a779ef721df3ba4e4540f449d76 /usr/local/bin/flux /usr/local/bin/flux
+COPY --from=quay.io/openbao/openbao:2.7.1@sha256:6d2b93856e3fcf7b18ad855a0b51eaba474dc8b79cf554379ea32034797d2acf /usr/bin/bao /usr/local/bin/bao
 COPY --from=gitea/tea:0.16.0@sha256:af50a2060ce4d8fd76b7ebe75a3d6db4bb0c615c5bc8f2bdfd8ea83207c4752d /usr/bin/tea /usr/local/bin/tea

 RUN talosctl version --client \
@@ -1,7 +1,7 @@
 # kuma-sync: reconciles KumaMonitor resources into Uptime Kuma. The unit tests
 # (no tags, so no Docker and no envtest) gate the image: the final stage copies
 # the binary out of this stage, so a failing test fails the build.
-FROM golang:1.27@sha256:2920bb124f1a2e12a1d39f0cab756ae0068441ee1da7655dec7f5e374944095b AS build
+FROM golang:1.27@sha256:f20d83a83e18dbd2a4714b8ffd4b30b287b6243962dd75608e86abcc63ba4efc AS build
 WORKDIR /src
 COPY go.mod go.sum ./
 RUN go mod download
@@ -9,7 +9,7 @@ COPY . .
 RUN go test ./...
 RUN CGO_ENABLED=0 go build -trimpath -ldflags="-s -w" -o /out/kuma-sync ./cmd/kuma-sync

-FROM gcr.io/distroless/static:nonroot@sha256:25ac8f0c3b2334a9eb05452266bd4697864a8c8ef0ef71e5e6480d7f92868329
+FROM gcr.io/distroless/static:nonroot@sha256:18f4cc26c9ffa7f20a359298be7deadac5c607c8e03a02b8119f1a5fc2f82c99
 COPY --from=build /out/kuma-sync /kuma-sync
 USER 65532
 ENTRYPOINT ["/kuma-sync"]
@@ -2,7 +2,7 @@
 # domains -> TLS passthrough -> Traefik ClusterIP) from Ingresses carrying the
 # public label. Standard library only, so the image is the interpreter plus one
 # file; runs unprivileged as its own user.
-FROM python:3.15-slim@sha256:94d3426f1def7db4a106a232ac121bfd975393e5dfd40a619b955fae45fe7913
+FROM python:3.15-slim@sha256:533b0b2f9b5d46fd097091cf826f27902f1d6c3bca714b6dc74551356aa97a51

 RUN useradd --system --no-create-home --shell /usr/sbin/nologin controller
 COPY controller.py /opt/controller.py
@@ -8,7 +8,7 @@ Issue #576. Five classes, each on its own VLAN.
 |---|---|---|
 | Ceph replication | `ceph-san`, VLAN 49 | the whole link is handed to the OSD pod (Multus `host-device`) |
 | pod↔pod | `pod0`, VLAN 48 | pod-sourced traffic is routed from table 47; bird puts the peers' pod CIDRs there via `pod0` |
-| world↔pod | `ingress0`, VLAN 47 | BGP session sourced there (next hop of every LoadBalancer VIP); pod egress takes table 47's default |
+| world↔pod | `ingress0`, VLAN 47 | BGP session sourced there (next hop of every LoadBalancer VIP); pod egress and VIP replies take table 47's default |
 | node↔node | `node0`, VLAN 46 | the node InternalIP and etcd live there; so do apiserver↔kubelet, cilium-health and host→remote-pod traffic |
 | world↔node | `mgmt0`, VLAN 40 | API VIP for kubectl, talosctl from outside, node egress (DNS, NTP, image pulls) |

@@ -22,13 +22,6 @@ Two rules, one per family, send every packet with a pod source to table 47
 (`talos/patches/site-homelab.yaml`, priorities 1000/1001, below main). The
 table decides by longest prefix; no rule carves out exceptions.

-A reply from a LoadBalancer VIP is pod traffic as well, but it reaches the
-routing decision with the VIP as its source. Two more rules (1002/1003) send
-the LoadBalancer pool to table 47, so the reply leaves over `ingress0`, where
-the request arrived as the BGP next hop. Without them main sent it out
-`mgmt0`. The pool is a site fact: the site patch's `lb-pool-site` ConfigMap,
-which `bgp-pool` takes by Flux substitution, and the rules beside it.
-
 bird (`kubernetes/infrastructure/bird/`) owns table 47 and fills it with:

 - the node's attached networks — `mgmt0`, `ingress0`, `pod0`, `node0` —
@@ -48,6 +41,15 @@ address, or a hostNetwork pod — would otherwise take the default to the UDM.
 The routing decision runs before Cilium's masquerade hook, so the rule sees the
 pod address and the masquerade then rewrites it to the egress link's.

+A reply from a LoadBalancer VIP is the exception: Cilium reverses the service
+NAT before the packet reaches the routing decision, so its source is the VIP
+and the pod rules miss it. bird's `vip-rules` container adds one rule per
+block of every `CiliumLoadBalancerIPPool` (priority 1002, `from <block>
+lookup 47`) and keeps them in sync with the pools every minute. The blocks
+must be CIDRs; a start/stop block stops the container. Without these rules
+the reply takes main's default over `mgmt0` while the request arrived on
+`ingress0` (#1547).
+
 Node-originated traffic does not match the rules and uses main, where Cilium
 routes each peer's pod CIDR via the peer's InternalIP — on `node0`.

@@ -48,7 +48,7 @@ spec:
        # cert when doing server-side OIDC token exchange. Bundle both
        # roots into one PEM since NODE_EXTRA_CA_CERTS accepts one file.
        - name: fetch-le-staging-roots
-          image: curlimages/curl:8.23.0@sha256:3fb7e13e5d179a845c97d78204f3f1064d1907037f26a3ebf91710571151fd93
+          image: curlimages/curl:8.23.0@sha256:11e64e28c5f58c8d93ee95fcf22ac055f48b0bc440e0ce7d8013bbb5ad6aa8ac
          command:
            - sh
            - -c
@@ -20,7 +20,7 @@ spec:
      # Pinned to match the samba-bridge fork's app image (server.Dockerfile
      # AUTHENTIK_VERSION). Bump in lockstep with the bridge rebuild — a
      # floating range drifts the chart ahead of the pinned app image.
-      version: "2026.8.5"
+      version: "2026.8.4"
      sourceRef:
        kind: HelmRepository
        name: authentik
@@ -28,7 +28,7 @@ spec:
      enableServiceLinks: false
      containers:
        - name: evcc
-          image: evcc/evcc:0.318.2@sha256:7ef9bcb15a4fdbfe3bf75d7a0857c40b6272b8783ea833fcbb22f8f134b336c7
+          image: evcc/evcc:0.318.1@sha256:2f8163ab22f04863afd86cd511c09607ba71debd9eebd65511f3d654922a80d1
          # evcc auto-discovers /etc/evcc.yaml on its default search path;
          # the file is delivered (with the MQTT password) via the
          # sops-encrypted evcc-config Secret mounted below.
@@ -88,7 +88,7 @@ spec:
        # deliver, and the ClusterIP has no backend while HA restarts. The
        # trap forwards SIGTERM so the pod still terminates promptly.
        - name: socat-ipv4
-          image: alpine/socat:1.8.1.3@sha256:466c8a0383e16463fabdb7df01008292fb78d07da60224a46cb5d3142ce122ed
+          image: alpine/socat:1.8.1.3@sha256:603174b77eebaba25ba8f8ebb15e1234d6600409554e95d6fb203fd106150287
          imagePullPolicy: IfNotPresent
          command:
            - /bin/sh
@@ -130,7 +130,7 @@ spec:
        # depends on the kernel's bindv6only default rather than on anything
        # stated here, so the roles are made explicit instead of inherited.
        - name: socat-ipv6
-          image: alpine/socat:1.8.1.3@sha256:466c8a0383e16463fabdb7df01008292fb78d07da60224a46cb5d3142ce122ed
+          image: alpine/socat:1.8.1.3@sha256:603174b77eebaba25ba8f8ebb15e1234d6600409554e95d6fb203fd106150287
          imagePullPolicy: IfNotPresent
          command:
            - /bin/sh
@@ -46,7 +46,7 @@ spec:
        k8s.v1.cni.cncf.io/networks: shelly-manager-iot
        # Changing this rolls the pod, which is the only way a change to the
        # shelly-manager-iot attachment takes effect.
-        net.renner.uno/iot-attachment-config: "tuning-no-ra-no-mtu-probing"
+        net.renner.uno/iot-attachment-config: "tuning-no-ra"
    spec:
      # eth0 (Cilium-managed pod network) is still the primary — it
      # carries the ingress, mosquitto LB return path, etc. The macvlan
@@ -64,7 +64,7 @@ spec:
          type: RuntimeDefault
      containers:
        - name: shelly-manager
-          image: gitea.renner.uno/konairius/shelly-manager:2026-11-02-1412 # {"$imagepolicy": "flux-system:shelly-manager"}
+          image: gitea.renner.uno/konairius/shelly-manager:2026-10-30-1343 # {"$imagepolicy": "flux-system:shelly-manager"}
          imagePullPolicy: IfNotPresent
          ports:
            - name: http
@@ -50,9 +50,6 @@ spec:
  # to IPv6 clients on v30iot leave via net1, bypassing Cilium's LoadBalancer
  # reverse NAT. tuning turns RA processing off on net1; the link-local
  # address mDNS uses stays.
-  #
-  # tcp_mtu_probing (pod-wide; Cilium sets 1): Shelly flash-write stalls
-  # trip black-hole detection, which cuts firmware downloads to 48-byte MSS.
  config: |
    {
      "cniVersion": "0.3.1",
@@ -70,8 +67,7 @@ spec:
          "type": "tuning",
          "sysctl": {
            "net.ipv6.conf.IFNAME.accept_ra": "0",
-            "net.ipv6.conf.IFNAME.autoconf": "0",
-            "net.ipv4.tcp_mtu_probing": "0"
+            "net.ipv6.conf.IFNAME.autoconf": "0"
          }
        },
        {
@@ -60,7 +60,7 @@ spec:
        # repo's configuration.yaml reads this key, so HA does not start
        # without it. Removing this init container takes the recorder down.
        - name: render-db-url
-          image: busybox:1.38@sha256:a46257e06c82ad85242d0d38077d0a52d9896389ca60a1a59d3463db586e5fa4
+          image: busybox:1.38@sha256:fd7dc98638c8e305f4dc34e979f1c0fdfdcaeb0fbf8fcff77ae834b6da3d7e6e
          command:
            - sh
            - -c
@@ -97,7 +97,7 @@ spec:
        # (homeassistant repo packages/matrix_assist.yaml), same filter-then-append
        # render as recorder_db_url above. Value is hex, nothing to escape.
        - name: render-matrix-password
-          image: busybox:1.38@sha256:a46257e06c82ad85242d0d38077d0a52d9896389ca60a1a59d3463db586e5fa4
+          image: busybox:1.38@sha256:fd7dc98638c8e305f4dc34e979f1c0fdfdcaeb0fbf8fcff77ae834b6da3d7e6e
          command:
            - sh
            - -c
@@ -275,7 +275,7 @@ spec:
        # any LAN client can edit, matching the user's trust-the-network
        # model. Runs as root so its writes match HA's root-owned files.
        - name: code-server
-          image: codercom/code-server:4.144.0@sha256:b9e8056bb74bd0c4f90284c7559eddeccd38400cfedc28e64c9d3ca50fef1f83
+          image: codercom/code-server:4.143.1@sha256:377b66bec5645510ed52549bb00e648c273edc604d230a3f90fd2b2433afcb97
          args:
            - "--auth=none"
            - "--bind-addr=0.0.0.0:8080"
@@ -337,7 +337,7 @@ spec:
        # dedicated long-lived token.
        # Runs as root to match HA's root-owned /config files. See config-sync.sh.
        - name: config-sync
-          image: alpine/git:2.54.0@sha256:ac6c828f15d82775d204d3d75112e1d34af8945da80ccf922713c9891ba68e7a
+          image: alpine/git:2.54.0@sha256:e8ceedc73bd80992ae8d58a7d24c58b8322146c354f1d2e81a97db4452068ec0
          command: ["/bin/sh", "/scripts/sync.sh"]
          env:
            - name: SYNC_BRANCH
@@ -36,7 +36,7 @@ spec:
          type: RuntimeDefault
      containers:
        - name: esphome
-          image: ghcr.io/esphome/esphome:2026.10.1@sha256:eab8cd6b4f51077c3883af4cedc058280ddc34a94f3b2ffee23e9d2cd5f1db64
+          image: ghcr.io/esphome/esphome:2026.10.0@sha256:1dff19a59070cdb523b76576cdc74d200f862d43c1407809088855920cf57145
          imagePullPolicy: IfNotPresent
          # Default entrypoint runs the dashboard (Device Builder) against the
          # mounted /config — no command override needed.
@@ -223,7 +223,7 @@ spec:
          type: RuntimeDefault
      containers:
        - name: controller
-          image: gitea.renner.uno/konairius/mqtt-account-controller:2026-11-02-1058 # {"$imagepolicy": "flux-system:mqtt-account-controller"}
+          image: gitea.renner.uno/konairius/mqtt-account-controller:2026-10-29-1634 # {"$imagepolicy": "flux-system:mqtt-account-controller"}
          env:
            - name: BROKER_URL
              value: ssl://mosquitto.home-assistant.svc.cluster.local:8884
@@ -163,7 +163,7 @@ spec:
          # Pinned tag. No Flux image-automation marker yet (unlike mosquitto) —
          # add an infrastructure/image-automation/zigbee2mqtt.yaml policy as a
          # follow-up if desired; pinned-and-manual is fine for the initial cutover.
-          image: ghcr.io/koenkk/zigbee2mqtt:2.14.3@sha256:fb4a6b0f6868ad1cd6a9996479802b2420e599c3463b62bbff8c8cd9a908f43a
+          image: ghcr.io/koenkk/zigbee2mqtt:2.14.2@sha256:addbfb08fcef2e4477de1535636118241e10d13e898fbe4d9faba8a4d7396311
          imagePullPolicy: IfNotPresent
          env:
            - name: TZ
@@ -17,7 +17,7 @@ spec:
    spec:
      containers:
        - name: ical-proxy
-          image: gitea.renner.uno/konairius/ical-proxy:2026-11-02-1055 # {"$imagepolicy": "flux-system:ical-proxy"}
+          image: gitea.renner.uno/konairius/ical-proxy:2026-10-28-0651 # {"$imagepolicy": "flux-system:ical-proxy"}
          ports:
            - containerPort: 8080
              name: http
@@ -66,7 +66,7 @@ spec:
        # The library PVC is populated by the seed phase before this
        # Deployment reconciles, so no rsync init is needed here.
        - name: fetch-le-staging-roots
-          image: curlimages/curl:8.23.0@sha256:3fb7e13e5d179a845c97d78204f3f1064d1907037f26a3ebf91710571151fd93
+          image: curlimages/curl:8.23.0@sha256:11e64e28c5f58c8d93ee95fcf22ac055f48b0bc440e0ce7d8013bbb5ad6aa8ac
          command:
            - sh
            - -c
@@ -57,7 +57,7 @@ spec:
        runAsNonRoot: false
      initContainers:
        - name: wait-for-influxd
-          image: influxdb:2.9.1@sha256:a450f6faf9573e6c2e96a0f575885ae8b7387cc87e4de94599e247ebfa3ac417
+          image: influxdb:2.9.1@sha256:8f106f7ff12659c5e367b99f8a0e7d14cbd4e6ca1f2914094062db6a9813d9c2
          env:
            - { name: INFLUX_HOST, value: http://influxdb-influxdb2.influxdb.svc.cluster.local:8086 }
            - { name: INFLUX_ORG, value: renner }
@@ -80,7 +80,7 @@ spec:
              echo "[$(date -u +%H:%M:%S)] $INFLUX_HOST ready and authenticated"
      containers:
        - name: create
-          image: influxdb:2.9.1@sha256:a450f6faf9573e6c2e96a0f575885ae8b7387cc87e4de94599e247ebfa3ac417
+          image: influxdb:2.9.1@sha256:8f106f7ff12659c5e367b99f8a0e7d14cbd4e6ca1f2914094062db6a9813d9c2
          env:
            - { name: INFLUX_HOST, value: http://influxdb-influxdb2.influxdb.svc.cluster.local:8086 }
            - { name: INFLUX_ORG, value: renner }
@@ -88,7 +88,7 @@ spec:
        # state). Two-stage so we don't try to create an auth before
        # the chart's `influx setup` has finished.
        - name: wait-for-influxd
-          image: influxdb:2.9.1@sha256:a450f6faf9573e6c2e96a0f575885ae8b7387cc87e4de94599e247ebfa3ac417
+          image: influxdb:2.9.1@sha256:8f106f7ff12659c5e367b99f8a0e7d14cbd4e6ca1f2914094062db6a9813d9c2
          env:
            - { name: INFLUX_HOST, value: http://influxdb-influxdb2.influxdb.svc.cluster.local:8086 }
            - { name: INFLUX_ORG, value: renner }
@@ -111,7 +111,7 @@ spec:
              echo "[$(date -u +%H:%M:%S)] $INFLUX_HOST ready and authenticated"
      containers:
        - name: create
-          image: influxdb:2.9.1@sha256:a450f6faf9573e6c2e96a0f575885ae8b7387cc87e4de94599e247ebfa3ac417
+          image: influxdb:2.9.1@sha256:8f106f7ff12659c5e367b99f8a0e7d14cbd4e6ca1f2914094062db6a9813d9c2
          env:
            - { name: INFLUX_HOST, value: http://influxdb-influxdb2.influxdb.svc.cluster.local:8086 }
            - { name: INFLUX_ORG, value: renner }
@@ -55,7 +55,7 @@ spec:
      repository: influxdb
      # Was pinned to VM 108's 2.7.4 for the one-time `influx restore`; that
      # migration is done. No downgrade below 2.9 without `influxd downgrade`.
-      tag: "2.9.1@sha256:a450f6faf9573e6c2e96a0f575885ae8b7387cc87e4de94599e247ebfa3ac417"
+      tag: "2.9.1@sha256:8f106f7ff12659c5e367b99f8a0e7d14cbd4e6ca1f2914094062db6a9813d9c2"
    # Single replica — InfluxDB OSS doesn't cluster. RWO Ceph-block + Velero
    # backups (pending) is the durability story.
    persistence:
@@ -48,7 +48,7 @@ spec:
            # this sentinel: the entrypoint copies AccountsSettings.json
            # into ephemeral /config-internal rather than symlinking it.
            - name: seed-config
-              image: busybox:1.38.0@sha256:a46257e06c82ad85242d0d38077d0a52d9896389ca60a1a59d3463db586e5fa4
+              image: busybox:1.38.0@sha256:fd7dc98638c8e305f4dc34e979f1c0fdfdcaeb0fbf8fcff77ae834b6da3d7e6e
              command: ["/bin/sh", "-eu", "-c"]
              args:
                - |
@@ -94,7 +94,7 @@ spec:
            - name: libation
              # Pinned by digest: Spegel caches mutable tag resolutions, so
              # :latest stayed on 13.7.11 despite imagePullPolicy Always.
-              image: docker.io/rmcrackan/libation:14.2.5@sha256:7bb47a417311d154c6691747320d0b62ce15527280ad99f0e2a85ac9a6b07e40
+              image: docker.io/rmcrackan/libation:14.2.2@sha256:fe188ac3c4ce63cc305971c76e648748341427e16bcf884011da8a384fd3908c
              # The entrypoint stages config into ephemeral /config-internal
              # by copy (only the DB is symlinked to the PVC), so Audible
              # tokens refreshed during a run are lost at container exit.
@@ -72,7 +72,7 @@ spec:
          type: RuntimeDefault
      containers:
        - name: element-web
-          image: ghcr.io/element-hq/element-web:v1.12.33@sha256:b0aa4b5a4fc1afaaff62e2dbd0c92d57e2b20724f3ccb2fef53aae949088c3d2
+          image: ghcr.io/element-hq/element-web:v1.12.32@sha256:f8700633d4336258af7a05b53865a4f479e0d628cf2d9ca83d28b5939314d4ce
          securityContext:
            allowPrivilegeEscalation: false
            readOnlyRootFilesystem: true
@@ -137,7 +137,7 @@ spec:
          type: RuntimeDefault
      containers:
        - name: mas
-          image: ghcr.io/element-hq/matrix-authentication-service:1.28.0@sha256:494c67f08ee8344368f4ed336611dea02b26c4e058bf37a212d090d50b612659
+          image: ghcr.io/element-hq/matrix-authentication-service:1.27.1@sha256:02712e68c90f3743fedbd5e3343517481b2b4f246a9fd9e2f050dc2553033847
          # The server applies database migrations and syncs the upstream
          # providers from the config on start.
          args: ["server", "--config", "/config/config.yaml"]
@@ -179,7 +179,7 @@ spec:
        # first. The marker on the data PVC makes later starts skip it.
        # Debug image: the plain one has no shell for the marker check.
        - name: migrate-to-mas
-          image: ghcr.io/element-hq/matrix-authentication-service:1.28.0-debug@sha256:362b8cc6b1176aec059f3539da2d47c8b47587d1584431bf1aca285146d42b08
+          image: ghcr.io/element-hq/matrix-authentication-service:1.27.1-debug@sha256:75f369b206785f1480f4a9f8dd81d7f9f3c6793c40fd184358ecd5d7acae17d4
          command:
            - /busybox/sh
            - -ec
@@ -234,7 +234,7 @@ spec:
              mountPath: /data
      containers:
        - name: synapse
-          image: ghcr.io/element-hq/synapse:v1.164.1@sha256:fa4aa980f279bcf26d4e7ecc2a2c9a701c73f2ebe8841da9bdcf3b2fdf693d75
+          image: ghcr.io/element-hq/synapse:v1.163.2@sha256:c3c21919a322f74d177a43bce8031adc81ef839ddd117a8dd03cfec350e76e02
          # Bypass /start.py: it expects to run as root and template a config
          # from env. Default mode is generate-missing-and-run, which creates
          # the signing key on first start.
@@ -17,7 +17,7 @@ spec:
  chart:
    spec:
      chart: kube-prometheus-stack
-      version: "92.17.0"
+      version: "92.13.1"
      sourceRef:
        kind: HelmRepository
        name: prometheus-community
@@ -74,7 +74,7 @@ spec:
          # alpine/kubectl is the maintained replacement. Pin the tag
          # to the cluster's server version so kubectl is in lock-step
          # (mismatched skew can break exec streaming).
-          image: docker.io/alpine/kubectl:1.37.2@sha256:5e2de75d37e811ba6eb92e06890b1b3263eae4f0a1c17b49269edec589b1352a
+          image: docker.io/alpine/kubectl:1.37.1@sha256:a5d0fbecab3f7b6c2ab7398f1daa73a3e0ed6ece18556d21698233e9e3759cc7
          env:
            - name: DATASET_TLV
              valueFrom:
@@ -47,7 +47,7 @@ spec:
        #
        # Hence socat is installed at startup and run inside this container.
        - name: otbr
-          image: openthread/otbr:latest@sha256:51f9af250252f4faae85d0197574a7c17ada873c78799854461a93e3e42cf324
+          image: openthread/otbr:latest@sha256:f3e43ccd86d6da87323873efa0c9b6f92a2a8cc75a804ea50e34c96680363e2e
          command:
            - bash
            - -c
@@ -287,7 +287,7 @@ spec:
              add: ["CHOWN", "DAC_OVERRIDE", "FOWNER", "SETGID", "SETUID"]
        # anchor: bridge-tls-proxy
        - name: bridge-tls-proxy
-          image: alpine/socat:1.8.1.3@sha256:466c8a0383e16463fabdb7df01008292fb78d07da60224a46cb5d3142ce122ed
+          image: alpine/socat:1.8.1.3@sha256:603174b77eebaba25ba8f8ebb15e1234d6600409554e95d6fb203fd106150287
          args:
            - "-d"
            - "TCP-LISTEN:11143,fork,reuseaddr,bind=127.0.0.1"
@@ -59,7 +59,7 @@ spec:
            - name: renovate
              # Pinned by tag and digest. Renovate updates itself through
              # homelab's renovate.json, like any other image here.
-              image: docker.io/renovate/renovate:44.199.2@sha256:0538f498be1b074e2434f8449cab7b2be30f8c0d0baf3b72814010f19db5bcfb
+              image: docker.io/renovate/renovate:44.181.8@sha256:b28b2de1dfd35de7e8198756d7f7cdd84ebc28c3f25e80f5b09e598d3faefb2a
              env:
                - name: RENOVATE_CONFIG_FILE
                  value: /config/config.json
@@ -32,7 +32,7 @@ spec:
          type: RuntimeDefault
      containers:
        - name: review-bot
-          image: gitea.renner.uno/gitea_admin/review-bot:2026-11-02-1054 # {"$imagepolicy": "flux-system:review-bot"}
+          image: gitea.renner.uno/gitea_admin/review-bot:2026-10-29-1346 # {"$imagepolicy": "flux-system:review-bot"}
          env:
            - name: GITEA_URL
              value: https://gitea.renner.uno
@@ -57,7 +57,7 @@ spec:
          # Pinned, and deliberately NOT under Flux image automation. rbac.yaml
          # is derived from what this exact version calls; an automatic bump
          # could add a call the Role does not cover, silently.
-          image: ghcr.io/stakater/ingressmonitorcontroller:v2.2.14@sha256:6d7b7fcb39d72296b6b6c0301b443fb31e12d86c55e706adcc7e4f92a887a3ba
+          image: ghcr.io/stakater/ingressmonitorcontroller:v2.2.13@sha256:3301afb61c10abc4e334deab9cd8601cef069e352b4d7f4dd0897787fca32eac
          command: ["/manager"]
          args:
            - --health-probe-bind-address=:8081
@@ -68,7 +68,7 @@ spec:
              type: RuntimeDefault
          containers:
            - name: sync
-              image: curlimages/curl:8.23.0@sha256:3fb7e13e5d179a845c97d78204f3f1064d1907037f26a3ebf91710571151fd93
+              image: curlimages/curl:8.22.0@sha256:58adaa4e8dca9c988bae2aba4ab3434a0bb2da16bbe3f92dec39ec7785166777
              command:
                - sh
                - -c
@@ -128,7 +128,7 @@ spec:
              type: RuntimeDefault
          containers:
            - name: run
-              image: curlimages/curl:8.23.0@sha256:3fb7e13e5d179a845c97d78204f3f1064d1907037f26a3ebf91710571151fd93
+              image: curlimages/curl:8.22.0@sha256:58adaa4e8dca9c988bae2aba4ab3434a0bb2da16bbe3f92dec39ec7785166777
              command:
                - sh
                - -c
@@ -25,7 +25,7 @@ spec:
        fsGroup: 33
      containers:
        - name: wordpress
-          image: wordpress:7.1.4-php8.5-apache@sha256:131532d893b5dc4fdf4312a878b7e5e4e4d1d40eb54c854c78eddf6282457dfb
+          image: wordpress:7.1.3-php8.5-apache@sha256:ce31904fa1e51b73543fa2e35cc8f1747589b2abf48dc747b422749b88b77e9e
          ports:
            - name: http
              containerPort: 80
@@ -89,9 +89,6 @@ spec:
    substituteFrom:
      - kind: ConfigMap
        name: cluster-settings
-      # The LoadBalancer pool, from the site's Talos patch (site-*.yaml).
-      - kind: ConfigMap
-        name: lb-pool-site
  dependsOn:
    - name: infrastructure
 ---
@@ -14,6 +14,10 @@ rules:
  - apiGroups: [""]
    resources: ["nodes"]
    verbs: ["get", "list"]
+  # vip-rules reads the LoadBalancer pool blocks it routes into table 47.
+  - apiGroups: ["cilium.io"]
+    resources: ["ciliumloadbalancerippools"]
+    verbs: ["get", "list"]
 ---
 apiVersion: rbac.authorization.k8s.io/v1
 kind: ClusterRoleBinding
@@ -69,7 +73,7 @@ spec:
          # first build of that day — imagePullPolicy: Always does not help,
          # because the tag itself resolves stale. Update this when the image
          # is rebuilt.
-          image: gitea.renner.uno/gitea_admin/bird@sha256:fee5645ad0d718661fb90da6ea9090d30fe47597bebeecb246c01c2adc6df585
+          image: gitea.renner.uno/gitea_admin/bird@sha256:65fc93efead1e9b065d2c78e748f6c3c1ad7216e28b757d9fa4e204fa2ffe9a9
          command: ["/bin/sh", "-c"]
          args:
            - |
@@ -95,7 +99,7 @@ spec:
              mountPath: /etc/bird
      containers:
        - name: bird
-          image: gitea.renner.uno/gitea_admin/bird@sha256:fee5645ad0d718661fb90da6ea9090d30fe47597bebeecb246c01c2adc6df585
+          image: gitea.renner.uno/gitea_admin/bird@sha256:65fc93efead1e9b065d2c78e748f6c3c1ad7216e28b757d9fa4e204fa2ffe9a9
          resources:
            requests:
              cpu: "50m"
@@ -111,6 +115,24 @@ spec:
          volumeMounts:
            - name: etc-bird
              mountPath: /etc/bird
+        # Routes replies from LoadBalancer VIPs into table 47 (vip-rules.sh).
+        - name: vip-rules
+          image: gitea.renner.uno/gitea_admin/bird@sha256:65fc93efead1e9b065d2c78e748f6c3c1ad7216e28b757d9fa4e204fa2ffe9a9
+          command: ["/bin/sh", "/config/vip-rules.sh"]
+          resources:
+            requests:
+              cpu: "5m"
+              memory: "16Mi"
+            limits:
+              cpu: "100m"
+              memory: "64Mi"
+          securityContext:
+            capabilities:
+              drop: ["ALL"]
+              add: ["NET_ADMIN"]
+          volumeMounts:
+            - name: config
+              mountPath: /config
      volumes:
        - name: config
          configMap:
@@ -10,3 +10,5 @@ configMapGenerator:
      - bird.conf
      - defines.tmpl
      - peers.tmpl
+      - vip-rules.sh
+      - vip-rules.tmpl
@@ -0,0 +1,66 @@
+#!/bin/sh
+# Keeps one routing rule per LoadBalancer pool block: from <block> lookup 47.
+#
+# A reply from a VIP reaches the routing decision with the VIP as its source:
+# Cilium reverses the service NAT before the packet enters the stack. The pod
+# rules (1000/1001, talos/patches/site-homelab.yaml) do not match it, so
+# without these rules it takes main's default over mgmt0 while its request
+# arrived on ingress0. The blocks are read from the pools, so the prefix is
+# written only there.
+#
+# Rules are left in place on exit: if bird is gone, table 47 is empty and the
+# lookup falls through to main, as it does for the pod rules.
+set -eu
+
+PRIO=1002
+TABLE=47
+
+family_of() {
+	case $1 in *:*) echo 6 ;; *) echo 4 ;; esac
+}
+
+reconcile() {
+	# An API error skips this round; the rules in place stay.
+	if ! blocks=$(kubectl get ciliumloadbalancerippools.cilium.io \
+		-o go-template-file=/config/vip-rules.tmpl); then
+		echo "reading the pools failed; retrying" >&2
+		return 0
+	fi
+	for b in $blocks; do
+		case $b in
+		range:*)
+			echo "pool block ${b#range:} is a range; only cidr blocks map to a rule" >&2
+			return 1
+			;;
+		esac
+	done
+
+	for fam in 4 6; do
+		want=$(for b in $blocks; do
+			if [ "$(family_of "$b")" = "$fam" ]; then echo "$b"; fi
+		done | sort -u)
+		have=$(ip -"$fam" rule show |
+			awk -v p="$PRIO:" -v t="$TABLE" '$1 == p && $2 == "from" && $4 == "lookup" && $5 == t { print $3 }' |
+			sort -u)
+
+		for b in $have; do
+			if ! echo "$want" | grep -qxF "$b"; then
+				ip -"$fam" rule del priority "$PRIO" from "$b" table "$TABLE"
+				echo "removed: priority $PRIO from $b table $TABLE"
+			fi
+		done
+		for b in $want; do
+			if ! echo "$have" | grep -qxF "$b"; then
+				ip -"$fam" rule add priority "$PRIO" from "$b" table "$TABLE"
+				echo "added: priority $PRIO from $b table $TABLE"
+			fi
+		done
+	done
+}
+
+trap 'exit 0' TERM INT
+while :; do
+	reconcile
+	sleep 60 &
+	wait $!
+done
@@ -0,0 +1,5 @@
+{{- /* One line per block of every enabled CiliumLoadBalancerIPPool. A start/stop
+     block renders as range:<start>-<stop>, which vip-rules.sh refuses. */ -}}
+{{- range .items }}{{ if not .spec.disabled }}{{ range .spec.blocks }}
+{{- if .cidr }}{{ .cidr }}{{ else }}range:{{ .start }}-{{ .stop }}{{ end }}
+{{ end }}{{ end }}{{ end -}}
@@ -24,7 +24,7 @@ spec:
      # window. The StatefulSet rolls one pod at a time (updateStrategyType
      # below). Undo after a pod has run the new server = restore the
      # snapshot, not a git revert.
-      version: ">=0.30.6 <0.31.0"
+      version: ">=0.30.4 <0.31.0"
      sourceRef:
        kind: HelmRepository
        name: openbao
@@ -71,7 +71,7 @@ spec:
              type: RuntimeDefault
          containers:
            - name: rotate
-              image: gitea.renner.uno/gitea_admin/secret-rotator:2026-11-02 # {"$imagepolicy": "flux-system:secret-rotator"}
+              image: gitea.renner.uno/gitea_admin/secret-rotator:2026-10-29 # {"$imagepolicy": "flux-system:secret-rotator"}
              args: ["ha-token"]
              securityContext:
                allowPrivilegeEscalation: false
@@ -133,7 +133,7 @@ spec:
              type: RuntimeDefault
          containers:
            - name: rotate
-              image: gitea.renner.uno/gitea_admin/secret-rotator:2026-11-02 # {"$imagepolicy": "flux-system:secret-rotator"}
+              image: gitea.renner.uno/gitea_admin/secret-rotator:2026-10-29 # {"$imagepolicy": "flux-system:secret-rotator"}
              args: ["ha-token"]
              securityContext:
                allowPrivilegeEscalation: false
@@ -195,7 +195,7 @@ spec:
              type: RuntimeDefault
          containers:
            - name: rotate
-              image: gitea.renner.uno/gitea_admin/secret-rotator:2026-11-02 # {"$imagepolicy": "flux-system:secret-rotator"}
+              image: gitea.renner.uno/gitea_admin/secret-rotator:2026-10-29 # {"$imagepolicy": "flux-system:secret-rotator"}
              args: ["ha-token"]
              securityContext:
                allowPrivilegeEscalation: false
@@ -45,7 +45,7 @@ spec:
              type: RuntimeDefault
          containers:
            - name: snapshot
-              image: quay.io/openbao/openbao:2.7.2@sha256:5046911075e137e154ce49ab051502f4ec7cf367d484ae4a4dce665d9b90561f # same image as the server — bao CLI included
+              image: quay.io/openbao/openbao:2.7.1@sha256:6d2b93856e3fcf7b18ad855a0b51eaba474dc8b79cf554379ea32034797d2acf # same image as the server — bao CLI included
              securityContext:
                allowPrivilegeEscalation: false
                capabilities:
@@ -8,12 +8,11 @@
 # Selector is opt-in via `lb-pool: bgp`. There is no default pool —
 # Services without this label do not get a LoadBalancer IP.
 #
-# The ranges are a site fact: lb-pool-site from the site's Talos patch
-# (talos/patches/site-*.yaml), by Flux substitution. A CIDR block uses all its
-# addresses (allowFirstLastIPs defaults to Yes).
-#
 # IPv4 + IPv6 ranges keep matching last bytes (10.2.99.x ↔ fd99::x)
 # so each LB has the same "tail" in both families.
+#
+# Blocks are CIDRs: bird's vip-rules turns each into a routing rule into
+# table 47, so VIP replies leave on ingress0 (infrastructure-inhouse/bird).
 apiVersion: cilium.io/v2alpha1
 kind: CiliumLoadBalancerIPPool
 metadata:
@@ -23,8 +22,8 @@ spec:
    matchLabels:
      lb-pool: bgp
  blocks:
-    - cidr: "${LB_POOL_V4}"
-    - cidr: "${LB_POOL_V6}"
+    - cidr: "10.2.99.0/24"
+    - cidr: "fd99::/120"
 ---
 # Cilium BGPv2 control-plane config peering each node with the UDM over the
 # ingress VLAN 47 — never over v30iot, which is the untrusted IoT segment.
@@ -67,18 +67,6 @@ cluster:
                nameservers:
                  - 185.12.64.1
                  - 185.12.64.2
-    # The LoadBalancer pool (see site-homelab.yaml). Services pin their
-    # addresses from the production pool, so a test run carries the same one.
-    - name: lb-pool-site
-      contents: |
-        apiVersion: v1
-        kind: ConfigMap
-        metadata:
-          name: lb-pool-site
-          namespace: flux-system
-        data:
-          LB_POOL_V4: 10.2.99.0/24
-          LB_POOL_V6: fd99::/120
 ---
 # Host DNS asks k8s-gateway first, so containerd resolves in-cluster names
 # (gitea.renner.uno) in the cluster; bootstrap.sh installs k8s-gateway right
@@ -85,18 +85,6 @@ cluster:
                nameservers:
                  - fd02:47::1
                  - 10.2.47.1
-    # The LoadBalancer pool (bgp-pool, infrastructure/post-crds/cilium-bgp.yaml,
-    # by Flux substitution). The routing rules 1002/1003 below name it too.
-    - name: lb-pool-site
-      contents: |
-        apiVersion: v1
-        kind: ConfigMap
-        metadata:
-          name: lb-pool-site
-          namespace: flux-system
-        data:
-          LB_POOL_V4: 10.2.99.0/24
-          LB_POOL_V6: fd99::/120
 ---
 # All pod traffic goes to table 47, and the table decides — longest prefix
 # match, no exceptions to carve out. bird fills it with the node's attached
@@ -114,7 +102,8 @@ cluster:
 #
 # Priorities are below main (32766) and clear of the rule Cilium installs at 9.
 # The two families are separate rule tables, so the priorities need not match;
-# the document name is the priority and must be unique.
+# the document name is the priority and must be unique. 1002 is taken by bird's
+# vip-rules: replies from LoadBalancer VIPs, whose source is the VIP.
 apiVersion: v1alpha1
 kind: RoutingRuleConfig
 name: "1000"
@@ -127,21 +116,6 @@ name: "1001"
 src: 10.244.0.0/16
 table: "47"
 ---
-# A LoadBalancer reply is pod traffic too, but it reaches the routing decision
-# with the VIP as its source, so the rules above miss it and main sends it out
-# mgmt0 while the request came in over ingress0. The pool is lb-pool-site above.
-apiVersion: v1alpha1
-kind: RoutingRuleConfig
-name: "1002"
-src: fd99::/120
-table: "47"
---
-apiVersion: v1alpha1
-kind: RoutingRuleConfig
-name: "1003"
-src: 10.2.99.0/24
-table: "47"
---
 # Host DNS asks k8s-gateway first, so containerd resolves in-cluster names
 # (gitea.renner.uno) in the cluster; bootstrap.sh installs k8s-gateway right
 # after the CNI, before anything else needs DNS. The UDM is the site's
@@ -47,15 +47,3 @@ cluster:
          values.yaml: |
            deployment:
              dnsPolicy: Default
-    # The LoadBalancer pool (see site-homelab.yaml). Services pin their
-    # addresses from the production pool, so a test run carries the same one.
-    - name: lb-pool-site
-      contents: |
-        apiVersion: v1
-        kind: ConfigMap
-        metadata:
-          name: lb-pool-site
-          namespace: flux-system
-        data:
-          LB_POOL_V4: 10.2.99.0/24
-          LB_POOL_V6: fd99::/120
@@ -10,8 +10,7 @@
    "dockerfile"
  ],
  "ignorePaths": [
-    "kubernetes/clusters/homelab/flux-system/**",
-    "kubernetes/infrastructure/image-automation/**"
+    "kubernetes/clusters/homelab/flux-system/**"
  ],
  "flux": {
    "managerFilePatterns": [
@@ -41,7 +40,6 @@
        "ghcr.io/home-assistant/home-assistant",
        "ghcr.io/immich-app/immich-machine-learning",
        "ghcr.io/immich-app/immich-server",
-        "ghcr.io/linuxserver/lidarr",
        "ghcr.io/matter-js/python-matter-server",
        "ghcr.io/music-assistant/server",
        "ghcr.io/paperless-ngx/paperless-ngx",
@@ -59,6 +57,7 @@
        "gitea.renner.uno/konairius/vsftpd",
        "gitea.renner.uno/konairius/wiki",
        "gotenberg/gotenberg",
+        "lscr.io/linuxserver/lidarr",
        "netbirdio/netbird",
        "quay.io/ceph/ceph",
        "redis",
@@ -2,7 +2,7 @@
 #
 # Nothing in the final image but the binary and the CA bundle it needs to reach the Mistral API
 # (the binary has no shell and no package manager to patch or to exec).
-FROM golang:1.27.3-bookworm@sha256:d9b6a1beb9474adc2a1ce842573639b44b2a0dd49b6669f51d3821baf9ec37c9 AS build
+FROM golang:1.27.3-bookworm@sha256:972b4110ba0f47539a975c31e52c64765e1ff19f2d74469ae4a595c6676e80a5 AS build

 WORKDIR /build
 COPY go.mod go.sum ./
@@ -8,7 +8,7 @@
 # lives inside the binary rather than in /etc/ssl. (In-cluster traffic to
 # OpenBao and Home Assistant is plain HTTP anyway; TLS is here so the same
 # binary can take over the external-API rotators — Cloudflare, Proxmox.)
-FROM rust:1.99-alpine@sha256:7e9e2d068a107a2adafd19930011178d14a78326053cd4c3a9f7ad3fd049557d AS build
+FROM rust:1.99-alpine@sha256:b1d5dd5a22a09404d3e798c0cf051b689077360374eb505fa5bb93c479693d87 AS build

 # musl-dev provides the C runtime bits the linker needs for a fully static
 # binary; nothing from it ends up in the final image.