feat(researchowl): daily SQLite backup CronJob + 5Gi backups PVC

Online-safe sqlite3 .backup at 03:00 Europe/Madrid, integrity check, 7-day retention. Data PVC mounted RW (WAL -shm requirement) but only read. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-04 10:43:41 +00:00
parent 214b4a10a4
commit ca520d3ca3
1 changed files with 90 additions and 0 deletions
@@ -0,0 +1,90 @@
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  annotations:
+    # Never let ArgoCD prune this PVC: it holds the DB backups.
+    argocd.argoproj.io/sync-options: Prune=false
+  name: researchowl-backups
+  namespace: researchowl
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      # DB is ~220MB and growing; 7 daily copies ≈ 1.6Gi, so 2Gi would fill up.
+      storage: 5Gi
+
+---
+apiVersion: batch/v1
+kind: CronJob
+metadata:
+  name: researchowl-db-backup
+  namespace: researchowl
+  labels:
+    app: researchowl
+spec:
+  schedule: "0 3 * * *"
+  timeZone: "Europe/Madrid"
+  concurrencyPolicy: Forbid
+  successfulJobsHistoryLimit: 3
+  failedJobsHistoryLimit: 3
+  jobTemplate:
+    spec:
+      backoffLimit: 2
+      activeDeadlineSeconds: 600
+      template:
+        metadata:
+          labels:
+            app: researchowl-db-backup
+        spec:
+          restartPolicy: Never
+          containers:
+            - name: backup
+              image: keinos/sqlite3:latest
+              command:
+                - /bin/sh
+                - -c
+                - |
+                  set -eu
+                  STAMP=$(date +%Y%m%d)
+                  DEST="/backups/researchowl-${STAMP}.db"
+                  echo "Backing up /data/researchowl.db -> ${DEST}"
+                  sqlite3 -cmd ".timeout 30000" /data/researchowl.db ".backup '${DEST}'"
+                  RESULT=$(sqlite3 "${DEST}" "PRAGMA integrity_check;")
+                  if [ "${RESULT}" != "ok" ]; then
+                    echo "Integrity check FAILED: ${RESULT}"
+                    rm -f "${DEST}"
+                    exit 1
+                  fi
+                  echo "Integrity check ok"
+                  # Retention: keep the 7 most recent backups
+                  for f in $(ls -1t /backups/researchowl-*.db | tail -n +8); do
+                    echo "Pruning old backup ${f}"
+                    rm -f "${f}"
+                  done
+                  ls -lh /backups/
+              volumeMounts:
+                # NOT readOnly: the source DB runs in WAL mode and sqlite readers
+                # must be able to touch the -shm/-wal sidecar files. The job only
+                # ever reads the DB (.backup); it never modifies application data.
+                - name: data
+                  mountPath: /data
+                - name: backups
+                  mountPath: /backups
+              resources:
+                requests:
+                  memory: "64Mi"
+                  cpu: "50m"
+                limits:
+                  memory: "256Mi"
+                  cpu: "500m"
+          volumes:
+            # Same RWO local-path PVC as the bot: the PV's node affinity pins this
+            # pod to the node where the data lives, so co-mounting is safe.
+            - name: data
+              persistentVolumeClaim:
+                claimName: researchowl-data
+            - name: backups
+              persistentVolumeClaim:
+                claimName: researchowl-backups