Creating Custom Images

The script allows you to build custom Amazon Machine Images (AMIs) or Azure VM images. You can use these images to create more instances of the original image. With a couple of exceptions, the steps for creating custom images are the same for AWS and Azure.

To create a custom image:

  1. Create the instance from which you want to build the image.
  2. Download or copy to the instance. For more information, see MapR Installer.
  3. Prep the image. This step adds MapR software to the image. For example:
    ./ image prep --version 5.2.2
  4. Respond to the questions in the script.
  5. (Optional): Install other software as needed on the image.
  6. (Azure only): Finalize the image:
    ./ image finalize azure
  7. Create the image by using the image create command:
    Note: (Azure only): The image create command must be run off the VM that you want to make an image out of. The command can be run anywhere the Azure CLI is installed or anywhere the MapR Installer 1.6 setup script is installed.
    • AWS example
      ./ image create
    • Azure example

      In this example, the script looks in the MapRImageCreateRG resource group for a machine called MyVirtualMachine and creates an image called MapRImage2:

      ./ image create azure -u -g MapRImageCreateRG -n MyVirtualMachine -i MapRImage2 
  8. (AWS only): Once the image is created, you can see the image ID in the EC2 console.
  9. (Azure only): Once the image is created, you can see the image in the list of resource groups. You can copy the image ID from the script:

    Note: Another way to obtain the Azure image ID is to navigate to the Azure console and check the resource ID for the new image:

  10. (AWS only): Update the AMI ID in the RegionMap of your CloudFormation template.
  11. (Azure only): Update the resource ID in the maprimageName field in the ARM template.

  12. (AWS only): Run the CloudFormation template using the steps in Running the MapR AWS Templates.
  13. (Azure only): Run the ARM template using the steps in Running the MapR Azure Templates.