Hashicorp Terraform is used to deploy AWS resources by writing code.
00-cis-hardening
- CIS Hardening for MDR root - (Ryan D'Amour, how does this go to other accounts)
00-organizations-and-iam
- IAM Roles and Policies across accounts (NOTE: No workspaces, applies everywhere)
00-state-mgmt
- S3 buckets for state management (may be prerequisite for others)
01-eips
- Elastic IPs and Associated DNS Record (protection from accidentally deletion)
02-msoc_vpc
- Managed SOC VPC (msoc is old name) - Meat and potatoes of command and control
03-mgmt
- ? Maybe Unused ? - Most appears to be junk, tread carefully.
04-ghe
- GitHub Enterprise - May be junk, GHE may be created elsewehere. Tread carefully.
05-customer_portal
- Web App for Customers in Docker using ECR, in its own vpc, running on ec2 running docker, not in fargate)
10-custpod1
- Splunk Monitoring Console + junk (Could probably burn and update)
11-codebuild
- Code Build to make RPMs
12-fargate
- Fargate for syslog-ng that gets ghe logs into moose
100-moose
- Our splunk environment (watch for modules of modules of modules)
101-afs
- AFS Customer Environment
102-saf
- SAF ("Smart and Final") - Powered Down through console - DO NOT TOUCH THE TF
103-nga
- FEDRAMP SPONSOR NGA ("National Gallery of Art"), sometimes referred to as Gallery.
104-coalfire
- Our FedRAMP Auditors (Standard customer with kali box)
105-cf2
- Our FedRAMP Auditors 2nd Environment
106-ma-c19
- Massachusetts Covid-19 (Internal AFS customer)
107-la-c19
- Louisiana Covid-19 (Internal AFS customer)
common
- Common files that are symbolicly linked into other folders
modules
- Reusable code - Do not run terraform here! A mix of homebrewed and third party modules.
https://warrensbox.github.io/terraform-switcher/
brew install warrensbox/tap/tfswitch
brew install warrensbox/tap/tgswitch
If there is a file that has a terraform version specified, running tfswitch
will automatically switch to that version.
06/2020
Enable debug export TF_LOG=DEBUG export TF_LOG_PATH=./terraform.log
Disable debug export TF_LOG=
05/2020
workspaces are being used to break up environments.
terraform workspace list terraform workspace select test
Strange errors? Unexpected results? try this rm .terraform terraform init
State issues terraform state show aws_ami.msoc_base terraform refresh -target=data.aws_ami.msoc_base
Terraform also has a DynamoDB State lock (msoc-terraform-lock). This will prevent terraform state breakage.
To manually remove the lock: https://www.terraform.io/docs/cli/commands/force-unlock.html
View TF code https://github.com/terraform-aws-modules
Modules
We are using the aws ec2-instance module
https://registry.terraform.io/modules/terraform-aws-modules/ec2-instance/aws/2.13.0 https://github.com/terraform-aws-modules/terraform-aws-ec2-instance
var.something means this is a module that needs the variable to run. Your code will fill the variable. data is a read-only terrafom object that queries provider or generates something on the localhost locals are variables that can refer to variables or other locals variables - expecting data from somewhere else. provider instance of the API
Some files are symlinks.
ln -s ../common/variables.tf variables.tf
ln -s ../amis.tf amis.tf
ln -s ../../../../prod/aws-us-gov/mdr-prod-c2/090-instance-vault/README.md README.md
IAM Role
get this error? aws_iam_policy.nga_instance_policy: Error creating IAM policy nga_instance_tag_read: AccessDenied:
add this provider = "aws.iam_admin"
in terraform .tf files when the self = true. that is for putting the security group into itself. e.g. add the security group to the security groups rules.
the terraform is setup in folders. each folder is a project and apply should be run in the folder. Common is the execption as some of the projects are dependent on that folder.
role and policy have to be done in the IAM terraform
iam_data.tf
02-msoc_vpc/lambda.tf with security groups
terraform plan -target= terraform plan -target=module.sensu_go_server.aws_instance.this -target=module.sensu_go_server.aws_route53_record.private
terraform apply -target=module.sensu_server.aws_route53_record.private -target=module.sensu_server.aws_instance.this
terraform apply -target=aws_security_group_rule.outbound_to_sensu -target=module.sensu_servers_sg.aws_security_group_rule.ingress_with_cidr_blocks[0] -target=module.sensu_servers_sg.aws_security_group_rule.ingress_with_cidr_blocks[1]
terraform apply -target=module.vpc_default_security_groups.aws_security_group_rule.typical_host_outbound_to_sensu_8081 -target=aws_security_group_rule.vault_server_to_sensu -target=module.vpc_default_security_groups.aws_security_group_rule.typical_host_outbound_to_sensu_5672
terraform apply -target=module.afs_cluster.module.vpc_default_security_groups.aws_security_group_rule.typical_host_outbound_to_sensu_5672 -target=module.afs_cluster.module.vpc_default_security_groups.aws_security_group_rule.typical_host_outbound_to_sensu_8081
We were generally using AWS 3 as soon as it came out. Now, the aws provider version is specified in xdr-terraform-live/terragrunt.hcl
.
Updating from TF12 to TF13 had no major issues. If you run into a module that isn't functioning first try:
terragrunt init --reconfigure
terragrunt apply
If it's an issue with a provider, you may need to add a required_providers tag to one of your files (e.g. providers-okta.tf
)that looks like this:
terraform {
required_providers {
okta = {
source = "oktadeveloper/okta"
}
}
}