cluster

commit 15cf535b78e7a293c488d5a5920f24701e56fe7d
parent 79be95afff297b1e0fdd408e4884f421122bac69
Author: MTRNord <MTRNord@users.noreply.github.com>
Date:   Fri,  3 Apr 2026 13:28:47 +0200

improve velero and pg backups

Signed-off-by: MTRNord <MTRNord@users.noreply.github.com>

Diffstat:
M
apps/talos_cluster/continuwuity/deployment.yaml
 | 
2
++
M
infrastructure_talos/configs/cnpg-cluster.yaml
 | 
2
+-
M
infrastructure_talos/configs/velero-schedules.yaml
 | 
4
++++
M
infrastructure_talos/controllers/velero/release.yaml
 | 
4
----
M
infrastructure_talos/controllers/velero/snapshot-class.yaml
 | 
2
+-

5 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/apps/talos_cluster/continuwuity/deployment.yaml b/apps/talos_cluster/continuwuity/deployment.yaml
@@ -14,6 +14,8 @@ spec:
       labels:
         app: continuwuity
       annotations:
+        # Explicit Kopia backup for this hcloud-volume (no CSI snapshots available)
+        backup.velero.io/backup-volumes: continuwuity-data
         # Freeze the process during Velero backup to ensure consistent RocksDB snapshot
         pre.hook.backup.velero.io/command: '["/bin/sh", "-c", "kill -STOP 1"]'
         pre.hook.backup.velero.io/timeout: "30s"
diff --git a/infrastructure_talos/configs/cnpg-cluster.yaml b/infrastructure_talos/configs/cnpg-cluster.yaml
@@ -333,7 +333,7 @@ metadata:
   name: hetzner-base-backup
   namespace: postgres-cluster
 spec:
-  schedule: 0 5 4 * * */3
+  schedule: 0 5 4 * * *
   backupOwnerReference: self
   immediate: true
   cluster:
diff --git a/infrastructure_talos/configs/velero-schedules.yaml b/infrastructure_talos/configs/velero-schedules.yaml
@@ -12,6 +12,7 @@ spec:
       - velero
       - kube-system
       - loki
+      - postgres-cluster
     storageLocation: default
     volumeSnapshotLocations:
       - default
@@ -31,6 +32,7 @@ spec:
       - velero
       - kube-system
       - loki
+      - postgres-cluster
     storageLocation: default
     volumeSnapshotLocations:
       - default
@@ -50,6 +52,7 @@ spec:
       - velero
       - kube-system
       - loki
+      - postgres-cluster
     storageLocation: default
     volumeSnapshotLocations:
       - default
@@ -69,6 +72,7 @@ spec:
       - velero
       - kube-system
       - loki
+      - postgres-cluster
     storageLocation: default
     volumeSnapshotLocations:
       - default
diff --git a/infrastructure_talos/controllers/velero/release.yaml b/infrastructure_talos/controllers/velero/release.yaml
@@ -50,10 +50,6 @@ spec:
       # when listing node-agent pods across many pods concurrently during backup.
       clientQPS: 100
       clientBurst: 200
-      # Use Kopia file-system backup for all volumes by default.
-      # CSI snapshots (for Longhorn) run additionally when a VolumeSnapshotClass
-      # labelled velero.io/csi-volumesnapshot-class=true exists.
-      defaultVolumesToFsBackup: true
       # Increase timeout to avoid "exposed PVB is not ready" errors on slow volumes
       podVolumeOperationTimeout: 60m
       backupStorageLocation:
diff --git a/infrastructure_talos/controllers/velero/snapshot-class.yaml b/infrastructure_talos/controllers/velero/snapshot-class.yaml
@@ -5,4 +5,4 @@ metadata:
   labels:
     velero.io/csi-volumesnapshot-class: "true"
 driver: driver.longhorn.io
-deletionPolicy: Delete
+deletionPolicy: Retain