Root cause of tonight's earlier CrowdSec/CNPG-backup workarounds: no node
in the cluster had any route into the Netbird mesh CIDR (100.108.0.0/16)
for pod-originated traffic. The per-namespace netbird 'router' pods are
inbound-only infrastructure (external peers reaching K8s services); their
own architecture has no reverse path.
Fix, in two parts:
1. infrastructure/netbird/manifests/egress-daemonset.yaml - one netbird
client per node, hostNetwork so its wt0 interface lives in the node's
real network namespace, plus a sidecar that adds a host route sending
100.108.0.0/16 out via it. hostNetwork requires a scoped Kyverno
PolicyException (infrastructure/kyverno/policies/netbird-egress-exception.yaml)
to the disallow-host-namespaces STIG policy - narrowly for this one
DaemonSet by name, not a namespace-wide exclusion.
2. Discovered the route alone wasn't enough for k3s NodePort traffic
(vps-minio:30900): Netbird manages its own nftables ACLs independent of
iptables/Kyverno, and its forward chain (netbird-rt-fwd) only permits
*established* connections through a peer acting as a router - never new
ones, by design, unless a Netbird 'Network Route' policy is explicitly
configured (it isn't, for this VPS). Locally-terminated connections
(tinyproxy) go through a separate, already-permissive ACL chain, which
is why the CrowdSec proxy fix from earlier tonight worked. Replicated
that working pattern for MinIO: minio-forward.service on the VPS host
(systemd, socat) forwards 100.108.113.41:9000 -> MinIO's ClusterIP,
avoiding the NodePort path entirely.
Re-enabled everything that was disabled/suspended earlier tonight because
of this gap, pointed at the new endpoint:
- CrowdSec CAPI/console-enroll (removed DISABLE_ONLINE_API, restored the
VPS proxy env vars)
- n8n/nextcloud/authentik CNPG backup.barmanObjectStore
- vault-raft-snapshot CronJob (unsuspended)
- nextcloud PVC content sync CronJob endpoint
vps-minio.netbird.internal is retired everywhere - it was never actually
resolvable (Netbird has no DNS configured) even before today's routing
fix.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
All 4 OSDs successfully recreated in clean raw mode (NVMe only, no DB devices).
Fixed talos-cp-02 NVMe device path from nvme0n1 to nvme1n1 to match actual hardware.
Removed temporary fix-rook-osd-metadata-device Kyverno policy as it's no longer needed.
Operator runs normally without manual patches.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
ArgoCD, NetBird operator, and Scylla Manager are upstream Helm charts
that don't set resource limits, probes, drop-all capabilities, or
disable SA token automount. These audit-mode policy violations inflated
the cumulative Prometheus failure counter to ~3800.
- argocd: exempt from require-resource-limits, require-probes
- netbird-operator: exempt from require-resource-limits,
disallow-automount-sa-token, require-probes
- scylla-manager: exempt from require-drop-all-capabilities,
disallow-automount-sa-token
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Use the OSD-specific activate-osd volume mount instead of rook-data
parent directory. Multiple old OSD dirs exist on each node and the
script was creating the symlink in the wrong (stale LVM) directory.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Use patchesJson6902 with append (/-) for create-block-db-symlink to
ensure it runs after activate, not before. The patchStrategicMerge was
inserting it before activate, causing prime-osd-dir to clear the symlink.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Remove metadataDevice (sdb) from all 4 nodes in cluster-values.yaml
since the SATA SSDs are slower than the NVMe data drives. Add temporary
Kyverno mutate policy to protect existing OSD deployments when the
operator is re-enabled - replaces expand-bluefs with a no-op and adds
conditional block.db symlink creation based on BlueStore label detection.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
