Run and destroy EC2 instance quickly

Script used to quickly create AWS VPC and launch there instance. Perfect when you need start temporary Instance for tests or some work.

Usage

Run ec2_manage script without arguments to see usage info.

Configuring

Export AWS credentials

export AWS_ACCESS_KEY_ID="YOUAWSACCESSKEY"
export AWS_SECRET_ACCESS_KEY="YoUrSecRetAcCesKeyHeRe"
export AWS_DEFAULT_REGION="us-east-1"

Edit these fields in settings file before you create AWS resources:

• aws_tag='temp-node' - used to tag all AWS resources created by script
• userdata='userdata.examples/user-data-file' - location of user-data script running on first boot
• ec2type='t2.medium' - EC2 type or running instance

NOTE: You can create local settings file named as settings.local that not synced with git repo and used to override any variables specified in settings. See settings.local.example

Examples

See usage info

$ ./ec2_manage

=== Dev AWS infra and Instance ===

Usage: ec2_manage <command>

Commands:

    help        display this help and exit

    noderun     launch  dev node
    nodedel     remove  dev node
    status      check node running
    ssh         ssh into running node

Create VPC and launch there EC2 instance

$ ./ec2_manage noderun

=== Dev AWS infra and Instance ===

Checking utils
Getting data about VPC
Getting data about instance
Adding AWS resources
Creating VPC
Tagging VPC
Creating Subnet
Modify Subnet for public IPs
Tagging Subnet
Creating  Internet Gateway
Tagging Gateway
Attaching Gateway to VPC
Obtaining Default Route table Id
Tagging Route Table
Adding default route
Attaching Subnet to Route Table
Creating Security Group
Tagging group
Adding open rule to group
Creating KeyPair
Launching t2.small Instance on AWS
Obtaining AMI if for Ubuntu 16.04
Checking userdata file present
Running instance
Tagging instance
Waiting while instance initalized
..........
Waitng for node
Checking ssh connection
.......
Connected
Waiting while orchestration finished
......................
Instance ready!
To login: ssh -i temp-node.pem [email protected]

Getting status of AWS resources

$ ./ec2_manage status

=== Dev AWS infra and Instance ===

Checking utils
Getting data about VPC
Getting data about instance
AWS VPC and correspoinding resoures created
Instance started
How to login: ssh -i temp-node.pem [email protected]

SSH into running instance

$ ./ec2_manage ssh

=== Dev AWS infra and Instance ===

Checking utils
Getting data about VPC
Getting data about instance
Welcome to Ubuntu 16.04.2 LTS (GNU/Linux 4.4.0-66-generic x86_64)

 * Documentation:  https://help.ubuntu.com
 * Management:     https://landscape.canonical.com
 * Support:        https://ubuntu.com/advantage

  Get cloud support with Ubuntu Advantage Cloud Guest:
    http://www.ubuntu.com/business/services/cloud

26 packages can be updated.
18 updates are security updates.


*** System restart required ***
Last login: Tue Mar 28 11:59:59 2017 from 46.4.69.7
To run a command as administrator (user "root"), use "sudo <command>".
See "man sudo_root" for details.

ubuntu@ip-10-222-222-123:~$

Destroy Instance and AWS VPC

$ ./ec2_manage nodedel

=== Dev AWS infra and Instance ===

Checking utils
Getting data about VPC
Getting data about instance
Deleting Instance
Waiting while Instance gone
..........................................
Instance removed
Removing AWS resources
Deleting Security Group
Deleting Subnet
Detaching Internet Gateway
Deleting Internet Gateway
Deleting VPC
Deleting KeyPair
AWS resources removed

User data scripts

Basic user-data script resides in userdata.examples/user-data-file. This script install docker engine on Instance. There are several examples of user-data script to use:

• userdata.examples/user-data-file.cuda - install Nvidia driver, docker, nvidia-docker plugin (EC2 type must be g2. or p2.)
• userdata.examples/user-data-file.mongo - install docker, run MongoDB server, configure authentication
• userdata.examples/user-data-file.k8s_single - install single-node Kubernetes cluster with components:
  • calico (pod network addon)
  • standalone heapster
  • nginx-ingresss-controller (with .spec.HostNetwork: 'true') - serve requests on port 80 of node
  • kubernetes-dashboard (accesible via http://<node_public_ip_address> with default credentials login: admin, password: single)

NOTE: If you using userdata.examples/user-data-file.k8s_single then please wait couple of minutes while all Kubernetes components started and than try login to Kubernetes dashboard from your browser

Notes

• Security Group created for Instance has rule to open all ports and all protocols to Internet
• On noderun script creates ssh pirivate key file placed next to script
• On nodedel ssh key file also deleted
• AWS resources script creates/removes:
  • AWS VPC with CIDR 10.222.0.0/16
  • Subnet with CIDR 10.222.222.0/24
  • Route Table
  • Association of Subnet with Route Table
  • Internet Gateway
  • Association Internet Gateway with VPC
  • Default route for Subnet to Internet Gateway
  • Security Group
  • Rule for Security group (simple rule full open to World)
  • KeyaPair and private SSH key (stored locally)
  • EC2 Instance with public IP address (see IP on ec2_manage status)