How do I resolve the "Server.InternalError: Internal error on launch" error for a failed stack in AWS CloudFormation?

Short description

You receive this error if duplicate or invalid device mappings are specified in your AWS CloudFormation template. You can't have two block devices map to the same location (for example, /dev/sdb).

Note: If you're using a Nitro-based instance type (for example, c5, m5, or t3), then you won't receive this error, because /dev/sdb and /dev/xvdb are mapped to two different NVMe devices in the operating system.

Resolution

In the BlockDeviceMappings property of your AWS CloudFormation template, confirm that your block devices aren't mapping to the same location by checking the value of DeviceName for each block device.

In the following JSON and YAML example templates, the block devices specified are /dev/xvdb and /dev/xvdc. The root volume is automatically provisioned for the instance, and the block devices are associated as secondary volumes.

JSON:

    "Ec2Instance" : {
      "Type" : "AWS::EC2::Instance", 
      "Properties" : {
        "...OtherProperties..."
        "BlockDeviceMappings" : [
          {
            "DeviceName" : "/dev/xvdb",
            "Ebs" : { "VolumeSize" : "100" }
          },{
            "DeviceName" : "/dev/xvdc",
            "Ebs" : { "VolumeSize" : "100" }
          }
        ]
      }
    }

YAML:

EC2Instance:
    Type: AWS::EC2::Instance
    Properties:
        ...OtherProperties...
        BlockDeviceMappings:
        -
          DeviceName: /dev/xvdb
          Ebs:
            VolumeSize: 100
        -
          DeviceName: /dev/xvdc
          Ebs:
            VolumeSize: 100

Related information

Block device mapping

EC2 block device mapping examples

Device names on Linux instances