Introduction

This image is used for automatic database backup on any S3 endpoint.

You can easily setup a cron on a kubernetes cluster for any kind of databases.

Image : skyloud/job-backup-db available on docker hub !

👉 https://hub.docker.com/r/skyloud/job-backup-db

Kind	Status
Postgres	✔ Done
Mysql	⚡️ In progress
MariaDB	✔ Done
MongoDB	✔ Done

You can help us to contribute on our repos 🚀

Other settings

How to follow backup jobs to be done with a heartbeat system. Use BAGCLI_HEARTBEAT_URL that will make a GET curl request at the end of job when succeed. Job will fail if request fail too.

Postgres

Configure S3

You'll need to setup a policy for a newly created user which will be used for this backup.

Copy this policy in postgres-prod-backup-policy.json :

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "s3:ListBucket"
      ],
      "Resource": [
        "arn:aws:s3:::project/",
        "arn:aws:s3:::project/*"
      ]
    },
    {
      "Effect": "Allow",
      "Action": [
        "s3:DeleteObject",
        "s3:GetObject",
        "s3:ListBucket",
        "s3:PutObject"
      ],
      "Resource": [
        "arn:aws:s3:::project/prod/postgres/*",
        "arn:aws:s3:::project/prod/postgres"
      ]
    }
  ]
}

Configure your policy on minio :

mc admin policy add s3 project-prod-postgres-policy s3-postgres-prod-policy.json
mc admin user add s3 project-prod-postgres-user my_very_strong_key_uuid_v4
mc admin policy set s3 project-prod-postgres-policy user=project-prod-postgres-user

MongoDB

You'll need to setup a policy for a newly created user which will be used for this backup.

Copy this policy in mongodb-prod-backup-policy.json :

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "s3:ListBucket"
      ],
      "Resource": [
        "arn:aws:s3:::project/",
        "arn:aws:s3:::project/*"
      ]
    },
    {
      "Effect": "Allow",
      "Action": [
        "s3:DeleteObject",
        "s3:GetObject",
        "s3:ListBucket",
        "s3:PutObject"
      ],
      "Resource": [
        "arn:aws:s3:::project/prod/mongodb/*",
        "arn:aws:s3:::project/prod/mongodb"
      ]
    }
  ]
}

Configure your policy on minio :

mc admin policy add s3 project-prod-mongodb-policy s3-mongodb-prod-policy.json
mc admin user add s3 project-prod-mongodb-user my_very_strong_key_uuid_v4
mc admin policy set s3 project-prod-mongodb-policy user=project-prod-mongodb-user

MariaDB

You'll need to setup a policy for a newly created user which will be used for this backup.

Copy this policy in mariadb-prod-backup-policy.json :

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "s3:ListBucket"
      ],
      "Resource": [
        "arn:aws:s3:::project/",
        "arn:aws:s3:::project/*"
      ]
    },
    {
      "Effect": "Allow",
      "Action": [
        "s3:DeleteObject",
        "s3:GetObject",
        "s3:ListBucket",
        "s3:PutObject"
      ],
      "Resource": [
        "arn:aws:s3:::project/prod/mariadb/*",
        "arn:aws:s3:::project/prod/mariadb"
      ]
    }
  ]
}

Configure your policy on minio :

mc admin policy add s3 project-prod-mariadb-policy s3-mariadb-prod-policy.json
mc admin user add s3 project-prod-mariadb-user my_very_strong_key_uuid_v4
mc admin policy set s3 project-prod-mariadb-policy user=project-prod-mariadb-user

How to use

RClone configuration

You'll need to configure rclone to access your S3 endpoint.

[backup]
type = s3
provider = minio
env_auth = false
access_key_id = my_very_strong_key_uuid_v4
secret_access_key = my_very_strong_key_uuid_v4
region = us-east-1
endpoint = https://s3.project.com
acl = private

With docker

Postgres Case

docker run --rm -it \
  -e BAGCLI_DATABASE_HOST="postgres.namespace.svc.cluster.local" \
  -e BAGCLI_DATABASE_USER="postgres" \
  -e BAGCLI_DATABASE_PASS="db_password" \
  -e BAGCLI_REMOTE_PATH="bucket/prod/postgres" \
  -e BAGCLI_WEBHOOK_URL=url \
  -e BAGCLI_WEBHOOK_CHANNEL: "#channel" \
  -v ./rclone.conf:/home/job/.config/rclone/rclone.conf \
  skyloud/job-backup-db postgres database-name

MongoDB case

docker run --rm -it \
  -e BAGCLI_DATABASE_URI="mongodb://mongoadmin:[email protected]:27017,mongodb-0.mongodb-headless.database:27017,mongodb-1.mongodb-headless.database:27017/mongodb_d_test?replicaSet=rs0&authSource=admin&retryWrites=true&w=majority" \
  -e BAGCLI_DATABASE_NAME="mongodb_d_test" \
  -e BAGCLI_REMOTE_PATH="bucket/prod/mongodb" \
  -e BAGCLI_WEBHOOK_URL=url \
  -e BAGCLI_WEBHOOK_CHANNEL: "#channel" \
  -v ./rclone.conf:/home/job/.config/rclone/rclone.conf \
  skyloud/job-backup-db mongodb

MariaDB case

docker run --rm -it \
  -e BAGCLI_MONGODB_HOST="mariadb.namespace.svc.cluster.local" \
  -e BAGCLI_MONGODB_PORT="3306" \
  -e BAGCLI_DATABASE_USER="test" \
  -e BAGCLI_DATABASE_PASS="test" \
  -e BAGCLI_REMOTE_PATH="bucket/prod/mariadb" \
  -e BAGCLI_WEBHOOK_URL=url \
  -e BAGCLI_WEBHOOK_CHANNEL: "#channel" \
  -v ./rclone.conf:/home/job/.config/rclone/rclone.conf \
  skyloud/job-backup-db mongodb

With Docker-compose

docker-compose build --nocache && docker-compose up -d

With kubernetes

PostgreSQL

Note : Please update the file kubernetes.yaml with your own values.

kubectl apply -f example/kubernetes_postgres.yaml

If you want to back up all the databases, you should take the example of this file :

kubectl apply -f example/kubernetes_postgres_dumpall.yaml

MongoDB

Note : Please update the file kubernetes_mongodb.yaml with your own values.

kubectl apply -f example/kubernetes_mongodb.yaml

MariaDB

Note : Please update the file kubernetes_mariadb.yaml with your own values.

kubectl apply -f example/kubernetes_mariadb.yaml

Notification

WebHook

If you want to set up a webhook call, you must define the following variables :

• BAGCLI_WEBHOOK_URL --> the complete url with its token
• BAGCLI_WEBHOOK_CHANNEL --> channel name

Enjoy !

ChangeLog

View our changelog here