Link Search Menu Expand Document

The objective here is to automate database backups with a cron job and keep the backups secure remotely on an s3 bucket.

  1. Installation:
    Install the backup gem on the machine. Versions other than 5.0.0.beta.1 versions can return an error with some json version. ssh onto your server and run gem install backup -v '5.0.0.beta.1'

  2. Configuration:
    Generate a model for the database. This model file will have the type and configuration of storage, database, encryption, compression, notifier etc. for the database.

    backup generate:model --trigger <model name> \ --databases="postgresql" --storages="s3" \ --compressor="gzip" --notifiers="slack"

    Refer this page for a detailed overview about the types of customizations available. You can omit any configuration that you don’t need here on the command above. This will create a /Users/<username>/Backup (macOS) directory by default (~/Backup on a linux machine). The config file can be found at /Users/<username>/Backup/models/<you-model-name>.rb or linux equivalent.<Project Environment>, <Project name with db type>) do
         database PostgreSQL do |db|               = ENV["POSTGRESQL_DATABASE"]
         db.username           = ENV["POSTGRESQL_USERNAME"]
         db.password           = ENV["POSTGRESQL_PASSWORD"]               = ENV["API_HOST_DOMAIN"]
         db.port               = 5432
         db.additional_options = ["-xc", "-E=utf8"]
      	store_with S3 do |s3|
         s3.access_key_id     = ENV["BACKUP_BOT_AWS_KEY"]
         s3.secret_access_key = ENV["BACKUP_BOT_AWS_SECRET"]
         s3.region            = "ap-south-1"
         s3.bucket            = <cm-live-projects bucket or storage provided by client>
         s3.path              = <Project name>
         s3.keep              = 5
         #notifier slack
      	notify_by Slack do |slack|
         slack.on_success = true
         slack.on_warning = true
         slack.on_failure = true
         slack.webhook_url = ENV["BACKUP_BOT_SLACK_WEBHOOK"]

    This is how a demo model file with mimimum complexity configuration looks like. All the variables in this particular file are being fetched from the OS’ environment variables (these were set through cloud66). There can be alternative ways to fetch these variables. The name, username and password of the database will be available on cloud66 (refer image below). A bucket was created to store these backups. A slack application was created to integrate a bot who sends the status of a backup.

    The config file can be edited in a terminal based text editor or (be copied to the server).


  3. Automation:
    To generate a backup use command backup perform -t <backup model name>. A cron job can be used for automatic backups. The command will differ in that case /usr/local/bin/backup perform -t staging here. With Ruby, use the whenever gem to schedule cron. Contents of crontab can be listed using crontab -l.

Current version of this gem does not have a restore command. As of now restoration has to be manual. An automated restoration mechanism will be available soon.