Kubegres

Kubegres allows to back-up the Primary PostgreSql database to a given volume in a scheduled way.

If the back-up feature is enabled in Kubegres YAML, behind the scenes a Kubernetes Cronjob resource is created which regularly executes a backup bash script. It is possible to override that bash script to have a better control of the backup process, as explained in the doc: Override the default configurations.

Create a PVC

If you do not have a PVC to store the contents to back-up, you can follow the example below.

Create a file:

vi my-backup-pvc.yaml

Add the following contents:

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: my-backup-pvc
  namespace: default
spec:
  storageClassName: "standard"
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 20Mi

Apply the changes:

kubectl apply -f my-backup-pvc.yaml

Enable the backup feature

Once the PVC is created, to enable the backup feature for a Postgres cluster, add the property "backup" to its Kubegres YAML, as in the example below:

backup:
    schedule: "0 */1 * * *"
    pvcName: my-backup-pvc
    volumeMount: /var/lib/backup

The above config has the following properties:

"backup.schedule: "0 */1 * * *" : it schedules a back-up every hour on the hour. Its format follows the standard Cron spec.
"backup.pvcName: my-backup-pvc" : the Persistence-Volume-Claim (PVC) defining where the back-up data will be stored. A volume backed by a PVC is mounted into the path defined by "backup.volumeMount" (see below).
"backup.volumeMount: /var/lib/backup" : the location where the backup is stored in the docker image. A volume is mounted on that path using the PVC name defined by "database.pvcName".

Please see below an example of a full Kubegres YAML config with those back-up properties:

apiVersion: kubegres.reactive-tech.io/v1
kind: Kubegres
metadata:
  name: mypostgres
  namespace: default

spec:

   replicas: 3
   image: postgres:16.1
   port: 5432

   database:
      size: 200Mi

   backup:
       schedule: "0 */1 * * *"
       pvcName: my-backup-pvc
       volumeMount: /var/lib/backup

   env:
      - name: POSTGRES_PASSWORD
        valueFrom:
           secretKeyRef:
              name: mySecretResource
              key: superUserPassword

      - name: POSTGRES_REPLICATION_PASSWORD
        valueFrom:
           secretKeyRef:
              name: mySecretResource
              key: replicationUserPassword

Check the backup is working

Once you applied the YAML changes above, you can monitor the CronJob as follows:

$ kubectl get CronJob

NAME                SCHEDULE      SUSPEND   ACTIVE   LAST SCHEDULE   AGE
backup-mypostgres   */1 * * * *   False     0        15s             14m

Each time a backup job is running, Kubernetes will create a Pod in the same namespace as your cluster of Postgres. You can check the logs of those pods to see if a backup task run successfully.

Example of backup jobs:

$ kubectl get pods

NAME                               READY   STATUS      RESTARTS   AGE
backup-mypostgres-27223836-25lzt   0/1     Completed   0          2m31s
backup-mypostgres-27223837-bshqr   0/1     Completed   0          91s
...

Checking the logs of a backup job:

$ kubectl logs backup-mypostgres-27223836-25lzt

05/10/2021 10:36:00 - Starting DB backup of Kubegres resource mypostgres into file: /var/lib/backup/mypostgres-backup-05_10_2021_10_36_00.gz
05/10/2021 10:36:00 - Running: pg_dumpall -h mypostgres-replica -U postgres -c | gzip > /var/lib/backup/mypostgres-backup-05_10_2021_10_36_00.gz
05/10/2021 10:36:00 - DB backup completed for Kubegres resource mypostgres into file: /var/lib/backup/mypostgres-backup-05_10_2021_10_36_00.gz

05 October 2021