Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/mcamacho97/terraform-mean-stack-aws/llms.txt

Use this file to discover all available pages before exploring further.

Running terraform destroy removes every AWS resource that Terraform created for this project in a single, coordinated operation. Terraform reads the current state file, builds a dependency graph of all managed resources, and tears them down in the correct reverse order. This includes the EC2 instances, load balancer, networking components, and IAM resources — the full stack is gone within a few minutes. Because this operation is irreversible for any data stored on those instances, take a moment to complete the pre-destroy checklist before proceeding.
terraform destroy is irreversible. EC2 instance storage is deleted permanently, and Elastic IP allocations are released back to the AWS pool. Ensure no critical application data or MongoDB documents exist on the stack before running the destroy command.

​
Pre-Destroy Checklist

1

Export any MongoDB data you need

The MongoDB instance has no automated backup. If you have stored data you want to keep, export it before destroying:
# SSH to the MongoDB instance (via jump host or SSM), then:
mongodump --out /tmp/backup
# Copy the dump to your local machine via scp or SSM file transfer
2

Back up the SSH key

The keys/<project_name>.pem file is written by a local_file Terraform resource, which means terraform destroy will delete it from disk. Back it up to a safe location before running destroy:
cp keys/terraform-mean.pem ~/secure-backups/terraform-mean.pem
3

Confirm the S3 backend bucket is out of scope

The S3 bucket that stores the Terraform state file was created outside this project and is not tracked in the state. It will not be destroyed. You can verify which bucket is configured in backend.tf before proceeding.

​
Run the Destroy Command

terraform destroy
Terraform prints a full destruction plan listing every resource it will remove, then prompts for confirmation: 
Do you really want to destroy all resources?
  Terraform will destroy all your managed infrastructure, as shown above.
  There is no undo. Only 'yes' will be accepted to confirm.

  Enter a value:
Type yes and press Enter. The operation typically takes 3–5 minutes, with the NAT Gateway and ALB taking the longest to deprovision. To skip the confirmation prompt in CI pipelines or scripted workflows: 
terraform destroy -auto-approve

​
What Gets Destroyed

All resources managed by this Terraform project are removed, including:

Compute

  • EC2 instance: node-1
  • EC2 instance: node-2
  • EC2 instance: mongodb
  • AWS Key Pair (<project_name>-key)
  • Local keys/<project_name>.pem file

Load Balancer

  • Application Load Balancer
  • Target group
  • HTTP listener
  • Target group attachments

Networking

  • VPC
  • Public subnets (×2)
  • Private subnet
  • Internet Gateway
  • NAT Gateway
  • Elastic IP (NAT)
  • Route tables and associations

Security & IAM

  • Security groups (ALB, Node, MongoDB)
  • IAM role (<project_name>-ec2-role)
  • IAM policy attachment (SSM)
  • IAM instance profile

​
What Is NOT Destroyed

ResourceReason
S3 state bucketCreated outside this project; not in Terraform state
Any MongoDB data stored on the instances is permanently lost when terraform destroy runs — EBS root volumes are deleted with their EC2 instances and are not recoverable. Export your data with mongodump before proceeding.

​
Partial Destroy

If you want to tear down only a specific module without removing the entire stack, use the -target flag: 
# Destroy only the Node 1 instance
terraform destroy -target=module.node_1

# Destroy only the load balancer
terraform destroy -target=module.alb
Partial destroys can leave the stack in an inconsistent state — for example, destroying the ALB while Node instances remain running. Use targeted destroy only for debugging or cost-saving experiments, not as a routine workflow.

​
Re-Deploying After Destroy

Once terraform destroy completes, the state file is empty and you can provision a completely fresh stack with: 
terraform apply
A new RSA 4096-bit SSH key pair is generated automatically and written to keys/<project_name>.pem. All IP addresses, DNS names, and instance IDs will differ from the previous deployment.
The NAT Gateway and Application Load Balancer account for the majority of the hourly cost in this stack. In development or lab environments, always run terraform destroy when the stack is not actively in use to avoid unnecessary charges — both resources accrue costs even when idle.

Build docs developers (and LLMs) love

Get started for freeTalk to us