AWS Ubuntu Desktop VM Module

Diagram

Overview

Generally, we have faced several issues around development environments with a diverse team. These include:

• Different underlying hardware causing duplication of effort
• Different underlying hardware having unique issues with software packages (particularly around M1 chips)
• Substantial bloatware around desktop environments causing several "bugs" isolated to individual developers
• Contributor individual hardware performance bottle necks

Several solutions exist to resolve one or more of these issues such as:

• Dev containers to replicate environments across developers
• VM software like Vagrant for creating repeatable VMs
• VSCode extensions committed along side source control

However, these have failed to meet our requirements. Bloatware and different underlying hardware issues persist with individuals having to solve software bugs unique to their hardware. Additionally, performance bottle necks can not be overcome by any software solution.

Due to this, we have frequently used Cloud Servers on AWS with custom built development environment AMIs to solve this issue.

This Terraform module aims to modularize our solution so that it may be reused by the OS community where a similar use case may arise.

Key Features

1. Spot instance requests with allow for cost effective instances.
2. SSH and RDP access limited to specific IP address ranges.
3. Automated backups of EBS volumes to allow for recovery

Base Images

This guide details how to create AMI images. We are currently developing some AMIs which will be backlogged however, these are not yet ready.

Ubuntu 22.04 with Chrome RDP and KDE Plasma 5

One image exists which launches Ubuntu 22.04 with Chrome RDP with KDE Plasma 5. This can be used as a base to get up and running and to begin installing custom software.

This was achieved by the following commands:

sudo apt update
sudo apt upgrade
sudo apt install --assume-yes wget tasksel


wget https://dl.google.com/linux/direct/chrome-remote-desktop_current_amd64.deb
sudo apt-get install --assume-yes ./chrome-remote-desktop_current_amd64.deb

sudo apt install tasksel
sudo apt install kde-plasma-desktop

// note the below line does not work and is done manually using nano at the moment.
sudo bash -c ‘echo “exec /etc/X11/Xsession /usr/bin/startplasma-x11” > /etc/chrome-remote-desktop-session
rm chrome-remote-desktop_current_amd64.deb

sudo reboot

Ubuntu 22.04 with Chrome RDP, KDE Plasma 5 and development software.

This configuration uses an AMI which was created by running the script.sh on an instance.

However, the AMI created uses an encrypted volume. So this can not be used for the module which uses a public AMI.

Set Up Chrome RDP

To set up Chrome RDP, go to the Chrome RDP Website and click on Set up via SSH.

From here click Begin, Next and Authorise.

Currently you must SSH into the instance to apply the command but we will provide a remote exec in a future release.

Features

Best Practices

Networking

• Creates a new independent VPC with a public subnet to remove dependency on pre-existing VPCs.

Security

• Spot instance security group only allows SSH, VNC and RDP inbound access from IP addresses specified by the user.
• A key pair is created using a public key provided by the user to allow authentication when accessing the instance using port 22(SSH).

Instance Configuration

• An elastic IP address is provisioned to preserve the public IP address of the instance between stops and starts. Note: The elastic IP address does not incur charges while the instance is running but will incur charges when stopped.

EBS Backup

• A Data Lifecycle Manager lifecycle policy is created which backups the instance daily before midnight and snapshots are retained for 1 week. In the event of catastrophic failure or loss of data, these snapshots can be used to restore EC2 instances. Instances to be managed by this policy are identified by the following tag: {AutomatedEbsBackups = true}.
• Snapshots are taken by the Data Lifecycle Manager.

Resource Material

Ubuntu Blog on Ubuntu Desktop on GCP
Guide to install KDE Plasma 5 Guide to generate SSH keys

Usage

module "ubuntu_desktop" {
    source = "../terraform_aws_ubuntu_desktop_vm"

    user_ip_address = "XXX"
    instance_ami = "XXX"

    tags = {
        Terraform   = "true"
        Environment = "dev"
  }
}

Requirements

Name	Version
aws	~>4.38.0

Providers

Name	Version
aws	4.38.0

Modules

Name	Source	Version
vpc	terraform-aws-modules/vpc/aws	~>3.18.1

Resources

Name	Type
aws_dlm_lifecycle_policy.instance_backup	resource
aws_eip.instance_ip	resource
aws_eip_association.instance_eip_assoc	resource
aws_iam_role.dlm_lifecycle_role	resource
aws_iam_role_policy.dlm_lifecycle	resource
aws_key_pair.instance_ssh_key	resource
aws_security_group.instance_sg	resource
aws_spot_instance_request.instance	resource
aws_region.current	data source

Inputs

Name	Description	Type	Default	Required
allowed_ingress_cidr_blocks	List of IP addresses allowed access to instance	list(string)	n/a	yes
associate_public_ip_address	Whether to associate a public IP address with an instance in a VPC	bool	true	no
environment	String to affix to any resource names and add to tags	string	"dev"	no
instance_ami	AMI to use for the instance	string	n/a	yes
instance_block_duration_minutes	The required duration for the Spot instances, in minutes. This value must be a multiple of 60 (60, 120, 180, 240, 300, or 360). The duration period starts as soon as your Spot instance receives its instance ID. At the end of the duration period, Amazon EC2 marks the Spot instance for termination and provides a Spot instance termination notice, which gives the instance a two-minute warning before it terminates. Note that you can't specify an Availability Zone group or a launch group if you specify a duration.	number	0	no
instance_interruption_behavior	Indicates Spot instance behavior when it is interrupted	string	"stop"	no
instance_monitoring	Whether detailed monitoring is enabled for EC2 instance	bool	true	no
instance_spot_price	The maximum price to request on the spot market	string	null	no
instance_type	Instance type to use for the instance.	string	"t3.xlarge"	no
prefix	string to prefix to resource names to make them easier to identify within the console	string	""	no
public_key	The public key material	string	n/a	yes
public_subnet_cidr_block	The IPv4 CIDR block of the public subnet	string	"10.0.0.0/28"	no
region	The region where the infrastructure will be deployed	string	"us-east-1"	no
root_block_device_delete_on_termination	Whether the volume should be destroyed on instance termination	bool	true	no
root_block_device_encryption	Whether to enable volume encryption	bool	true	no
root_block_device_size	Size of the volume in gibibytes	number	32	no
root_block_device_throughput	Throughput to provision for a volume in mebibytes per second (MiB/s)	number	125	no
root_block_device_type	The type of volume	string	"gp3"	no
tags	Tags to add to all created resources	map(any)	{}	no
vpc_cidr_block	The IPv4 CIDR block of the VPC	string	"10.0.0.0/26"	no

Outputs

Name	Description
elastic_ip_id	The ID of the elastic IP
instance_arn	The arn of the instance
instance_id	The ID of the instance
instance_private_ip	The private ip address of the instance
instance_public_ip	The public ip address of the instance
instance_sg_arn	The arn of the instance security group
instance_sg_id	The ID of the instance security group
instance_ssh_key_arn	The arn of the instance ssh key pair
instance_ssh_key_id	The ID of the instance ssh key pair
lifecycle_policy_arn	The arn of the lifecycle policy
lifecycle_policy_id	The arn of the lifecycle policy
public_subnet_arn	The public subnet arn
public_subnet_cidr_block	The public subnet CIDR block
public_subnet_id	The public subnet ID
vpc_arn	The arn of the VPC
vpc_id	The ID of the VPC

Contributing

Bug Reports & Feature Requests

Please use the issue tracker to report any bugs or file feature requests.

Contributors

Name	Role
Muhammad Hasan	Lead Developer
Faizan Raza	Developer

Style Guidelines

Naming Convention

Throughout the module, the naming convention used is as follows:

• For Terraform resource names, the names consist of lowercase characters and _ to represent spaces in the name.
• For AWS resource names, the names consist of lowercase characters and - to represent spaces in the name, with the prefix variable attached to the beginning of the name and the environment.

Resources Used

Terraform Docs used for generating documentation.