Merge pull request 'cnpg: cap stopDelay to bound the barman-plugin failover wedge' (#18) from fix/cnpg-stopdelay-wedge-mitigation into master

Reviewed-on: gitea_admin/homelab#18

kubernetes/seed/authentik/cluster.yaml

@@ -12,6 +12,13 @@ spec:
   # -rw backendless that whole time. Capping at 10s bounds failover downtime to
   # ~10-15s (fast shutdown + promotion).
   smartShutdownTimeout: 10
+  # Bound the post-failover wedge: with the barman-cloud plugin the demoted
+  # primary's instance-manager deadlocks on shutdown (the native sidecar never
+  # terminates — upstream cloudnative-pg#9770), so the pod hangs Terminating
+  # until SIGKILL at terminationGracePeriodSeconds (= stopDelay). The default
+  # 1800s leaves the cluster at 1/2 for ~30min after every failover; 60s caps
+  # that to ~1min. Postgres itself is already cleanly down by ~12s.
+  stopDelay: 60
   imageName: ghcr.io/cloudnative-pg/postgresql:16
   enableSuperuserAccess: true
   storage:

kubernetes/seed/immich/cluster.yaml

@@ -9,6 +9,12 @@ spec:
   # paths, blocking replica join. Bump back to 2 once Immich's in-app
   # vchord migration drops the vectors extension.
   instances: 1
+  # See ak-pg cluster.yaml for rationale. Single-instance, so no failover, but
+  # the barman-plugin shutdown wedge (cloudnative-pg#9770) takes the whole DB
+  # down until SIGKILL — bounding terminationGracePeriod (stopDelay) matters
+  # even without a standby. smartShutdownTimeout keeps pod restarts quick.
+  smartShutdownTimeout: 10
+  stopDelay: 60
   imageName: ghcr.io/corentingiraud/cnpg-pgvector-vectorchord:16-migration
   enableSuperuserAccess: true
   storage:

kubernetes/seed/paperless/cluster.yaml

@@ -5,6 +5,11 @@ metadata:
   namespace: paperless
 spec:
   instances: 2
+  # See ak-pg cluster.yaml for rationale: smartShutdownTimeout bounds failover
+  # downtime (the -rw backendless window); stopDelay bounds the barman-plugin
+  # shutdown-wedge window (cloudnative-pg#9770) from ~30min to ~1min.
+  smartShutdownTimeout: 10
+  stopDelay: 60
   imageName: ghcr.io/cloudnative-pg/postgresql:16
   enableSuperuserAccess: true
   storage: