docker-pgdump-gcs

A simple Docker container to perform a pg_dump command and upload the archive file to GCS. You can specify env variables used by the pg_dump to customize the execution (PGDATABASE, PGHOST, PGPORT, etc.).

This container is meant to be used as a cronjob with Kubernetes, and it will simply dump the data as a plain SQL file with compression enabled.

The service account running the container needs to have write access to the destination bucket, as this container makes use of gsutil.

Volumes

You can mount a volume on /backups if you want to collect the intermediate archives, or give more space to your container for the backup process.

Cron Schedule Syntax

# ┌───────────── minute (0 - 59)
# │ ┌───────────── hour (0 - 23)
# │ │ ┌───────────── day of the month (1 - 31)
# │ │ │ ┌───────────── month (1 - 12)
# │ │ │ │ ┌───────────── day of the week (0 - 6) (Sunday to Saturday;
# │ │ │ │ │                                   7 is also Sunday on some systems)
# │ │ │ │ │
# │ │ │ │ │
# * * * * *

Please refer to https://kubernetes.io/docs/concepts/workloads/controllers/cron-jobs/#cron-schedule-syntax

Kubernetes definition example

apiVersion: batch/v1
kind: CronJob
metadata:
  name: postgresql-backup
spec:
  schedule: "0 */6 * * *"
  jobTemplate:
    spec:
      template:
        spec:
          containers:
          - name: pgdump-gcs-cron
            image: openrm/pgdump-gcs:latest
            imagePullPolicy: IfNotPresent
            env:
            - name: PGDATABASE
              value: "postgresql_database"
            - name: PGHOST
              value: "postgresql_host"
            - name: PGPORT
              value: "postgresql_port"
            - name: PGUSER
              valueFrom:
                secretKeyRef:
                  name: postgresql
                  key: user
            - name: PGPASS
              valueFrom:
                secretKeyRef:
                  name: postgresql
                  key: password
            - name: GCS_BUCKET
              value: "backup_bucket_name"

In this simple example, the cron is started every 6 hours.

Terraform

resource "kubernetes_cron_job_v1" "postgresql_backup" {
  metadata {
    name = "postgresql-backup"
  }
  spec {
    schedule                  = "0 */6 * * *"
    job_template {
      spec {
        template {
          spec {
            container {
              name            = "pgdump-gcs-cron"
              image           = "openrm/pgdump-gcs:latest"
              env {
                name = "PGDATABASE"
                value = postgresql_database
              }
              env {
                name = "PGHOST"
                value = postgresql_host
              }
              env {
                name = "PGPORT"
                value = postgresql_port
              }
              env {
                name = "PGUSER"
                value_from {
                  secret_key_ref {
                    name = "postgresql"
                    key  = "user"
                  }
                }
              }
              env {
                name = "PGPASS"
                value_from {
                  secret_key_ref {
                    name = "postgresql"
                    key  = "password"
                  }
                }
              }
              env {
                name = "GCS_BUCKET"
                value = "backup_bucket_name"
              }
            }
          }
        }
      }
    }
  }
}

In this simple example, the cron is started every 6 hours.