Contents

Empty

Similar to Docker, a data image is created based on a definition contained in a .yaml file. To get started quickly, we will create a simple empty data image which will allow us to create an empty SQL Server 2019 data container.

Create a file development.yaml with your data image specifications using the empty source for a SQL Server 2019 database engine:

sourceType: empty
name: dev
engine: mssql
version: 2019

And here's an equivalent example but for PostgreSQL 14:

sourceType: empty
name: empty-pg
engine: postgresql
version: 14

Likewise, one for MySQL 8.0:

sourceType: empty
name: empty-mysql
engine: mysql
version: 8

You can now jump to Creating a data image (or alternatively, choose a different source type below for your data image definition).

Backups

Another option is to provide a database backup file as a source for a data image:

sourceType: Backup
name: backup-mssql
engine: mssql
version: 2019
backups:
  - path: my-backups/backup1.bak

The backup will be processed from a file share. The file path has to be provided relative to the configured file share root. The entire file share will be mounted to the temporary database to perform the backup restore

You can restore multiple backup files by adding multiple entries in the backups section.

sourceType: Backup
name: backup-mssql
engine: mssql
version: 2019
backups:
  - path: my-backups/backup1.bak
  - path: my-backups/backup2.bak

In this example we want to create an image by restoring two backup files: backup1.bak and backup2.bak.

For PostgreSQL custom backups, please see the examples on this additional page.

Striped Backups

For SQL Server, Redgate Clone supports striped backup files. This feature is currently gated behind a feature flag.

To use the feature, place all the backup files into a single directory on the share and then provide the path to the directory as the path . Redgate Clone will use all the files in this directory as part of a single restore operation in SQL Server.

To turn on this functionality, issue the following command from a machine that has kubectl access to the cluster and the kots plugin installed (for instance, one of the cluster nodes if using an Embedded installation)

kubectl kots set config cloning-capability-app --key multifile_backups_enabled --value "1" --deploy

To turn the feature off after enabling it, issue the following command

kubectl kots set config cloning-capability-app --key multifile_backups_enabled --value "0" --deploy

Pre and post scripts

You can run a SQL script as a part of data image creation process to modify the databases included in the image. The scripts can be specified as a pre- or a postScript, allowing to make modifications both before and after an empty database is created (for empty images) or a backup is restored (for data images created from backups). For example, a script can be used to add tables and/or data to an empty data image, or to modify user permissions on a data image restored from a backup. These changes is stored in the data image and will be included into all data containers created from it.

To run a pre-script:

sourceType: Backup
name: backup-mssql
engine: mssql
version: 2019
backups:
  - path: my-backups/backup1.bak
preScript: |
  CREATE DATABASE [myDatabase]

In this example, the pre-script creates an empty database called "myDatabase" before restoring the backup file.

Likewise, you can also add a post-script to perform SQL operations after the database backup has been restored (or the empty data container has been created) (e.g. set permissions). Here's an example:

sourceType: Backup
name: backup-mssql
engine: mssql
version: 2019
backups:
  - path: bak/AdventureWorksLT2019.bak
preScript: |
  CREATE DATABASE prescript
postScript: |
  CREATE TABLE [AdventureWorksLT2019].[dbo].[Sheep] (id INT)

Note: If your pre or post script is multi-line, make sure you indent it or you may get an error about " Operation failed (OperationFailed=41):" and "could not find expected ':'"

The order of operations during data image creation will be as follows:

• An empty SQL Server instance is spin up.
• Database prescript is created (pre-script).
• Database AdventureWorksLT2019 is created (database restore).
• Table dbo.Sheep is added to the AdventureWorksLT2019 database (post-script).

And the outcome will be: