How to fix Error: Error creating CodePipeline: PipelineNameInUseException in Terraform

First, verify that the pipeline exists in AWS CodePipeline:

aws codepipeline list-pipelines --region us-east-1

Look for your pipeline name in the output. You can also check the AWS Console by navigating to CodePipeline and searching for the pipeline name.

If the pipeline exists, proceed to the next step. If it doesn't exist, check for typos in your Terraform configuration.

If the pipeline already exists in AWS and you want Terraform to manage it, import it into your Terraform state:

terraform import aws_codepipeline.your_resource_name your-pipeline-name

Replace your_resource_name with the logical name of your aws_codepipeline resource in Terraform (the part after the dot in your .tf file).

Example:

terraform import aws_codepipeline.main my-ci-pipeline

After a successful import, Terraform will update your state file to track the existing pipeline. Verify the import:

terraform state show aws_codepipeline.your_resource_name

If the existing pipeline is orphaned or no longer needed, delete it from AWS:

aws codepipeline delete-pipeline --name your-pipeline-name --region us-east-1

Warning: Deleting a pipeline also deletes its history and any stored artifacts. Be sure you don't need it before proceeding.

After deletion, Terraform can create the new pipeline:

terraform plan
terraform apply

If you want both pipelines to coexist, rename your pipeline in the Terraform configuration:

resource "aws_codepipeline" "main" {
  name     = "my-new-pipeline-name"  # Use a unique name
  role_arn = aws_iam_role.codepipeline.arn

  # ... rest of configuration
}

Then apply the changes:

terraform plan
terraform apply

Make sure the new name is unique within your AWS account and region.

To avoid future state sync issues, verify your Terraform state matches AWS reality:

terraform refresh

This command updates your local state to reflect the current state of resources in AWS without making any changes.

Then check if there are any differences:

terraform plan

If terraform plan shows no changes, your state is consistent with AWS.

To prevent future state loss, store your Terraform state in a remote backend rather than locally:

For AWS S3 + DynamoDB:

terraform {
  backend "s3" {
    bucket         = "my-terraform-state"
    key            = "codepipeline/terraform.tfstate"
    region         = "us-east-1"
    dynamodb_table = "terraform-locks"
    encrypt        = true
  }
}

For Terraform Cloud:

terraform {
  cloud {
    organization = "my-org"

    workspaces {
      name = "production"
    }
  }
}

A remote backend ensures your state is durable, shared across team members, and protected from accidental deletion.