Módulo de Terraform para AWS Backup
Acabo de publicar mi primer módulo Terraform de código abierto llamado terraform-aws-backup en Github, así que decidí compartirlo también en el Registro público de Terraform.
Puedes verificar mi módulo terraform-aws-backup en el Terraform Registry o clonarlo de Github
Si quieres echar un vistazo al módulo, también dejé el archivo README en esta publicación:
terraform-aws-backup
Terraform module to create AWS Backup plans. AWS Backup is a fully managed backup service that makes it easy to centralize and automate the back up of data across AWS services (EBS volumes, RDS databases, DynamoDB tables, EFS file systems, and Storage Gateway volumes).
Usage
You can use this module to create a simple plan using the module’s rule_*
variables. You can also use the rules
and selections
list of maps variables to build a more complete plan by defining several rules and selections at once.
Check the examples for the simple plan, the simple plan with list and the complete plan snippets.
Example (complete plan)
This example creates a plan with two rules and two selections at once. It also defines a vault key which is used by the first rule because no target_vault_name
was given (null). Whereas the second rule is using the “Default” vault key.
The first selection has two assignments, the first defined by a resource ARN and the second one defined by a tag condition. The second selection has just one assignment defined by a resource ARN.
module "aws_backup_example" {
source = "../modules/terraform-aws-backup"
# Vault
vault_name = "vault-3"
# Plan
plan_name = "complete-plan"
# Multiple rules using a list of maps
rules = [
{
name = "rule-1"
schedule = "cron(0 12 * * ? *)"
target_vault_name = null
start_window = 120
completion_window = 360
lifecycle = {
cold_storage_after = 0
delete_after = 90
},
recovery_point_tags = {
Environment = "production"
}
},
{
name = "rule-2"
target_vault_name = "Default"
schedule = null
start_window = 120
completion_window = 360
lifecycle = {}
recovery_point_tags = {}
},
]
# Multiple selections
# - Selection-1: By resources and tag
# - Selection-2: Only by resources
selections = [
{
name = "selection-1"
resources = ["arn:aws:dynamodb:us-east-1:123456789101:table/mydynamodb-table1"]
selection_tag = {
type = "STRINGEQUALS"
key = "Environment"
value = "production"
}
},
{
name = "selection-2"
resources = ["arn:aws:dynamodb:us-east-1:123456789101:table/mydynamodb-table2"]
selection_tag = {}
},
]
tags = {
Owner = "backup team"
Environment = "production"
Terraform = true
}
}
Inputs
Name | Description | Type | Default | Required |
---|---|---|---|---|
plan_name | The display name of a backup plan | `string` | n/a | yes |
rule_completion_window | The amount of time AWS Backup attempts a backup before canceling the job and returning an error | `number` | n/a | yes |
rule_lifecycle_cold_storage_after | Specifies the number of days after creation that a recovery point is moved to cold storage | `number` | n/a | yes |
rule_lifecycle_delete_after | Specifies the number of days after creation that a recovery point is deleted. Must be 90 days greater than `cold_storage_after` | `number` | n/a | yes |
rule_name | An display name for a backup rule | `string` | n/a | yes |
rule_recovery_point_tags | Metadata that you can assign to help organize the resources that you create | `map(string)` | `{}` | no |
rule_schedule | A CRON expression specifying when AWS Backup initiates a backup job | `string` | n/a | yes |
rule_start_window | The amount of time in minutes before beginning a backup | `number` | n/a | yes |
rules | A list of rule maps | `list` | `[]` | no |
selection_name | The display name of a resource selection document | `string` | n/a | yes |
selection_resources | An array of strings that either contain Amazon Resource Names (ARNs) or match patterns of resources to assign to a backup plan | `list` | `[]` | no |
selection_tag_key | The key in a key-value pair | `string` | n/a | yes |
selection_tag_type | An operation, such as StringEquals, that is applied to a key-value pair used to filter resources in a selection | `string` | n/a | yes |
selection_tag_value | The value in a key-value pair | `string` | n/a | yes |
selections | A list of selction maps | `list` | `[]` | no |
tags | A mapping of tags to assign to the resource | `map(string)` | `{}` | no |
vault_kms_key_arn | The server-side encryption key that is used to protect your backups | `string` | n/a | yes |
vault_name | Name of the backup vault to create. If not given, AWS use default | `string` | n/a | yes |
Outputs
Name | Description |
---|---|
plan_arn | The ARN of the backup plan |
plan_id | The id of the backup plan |
plan_version | Unique, randomly generated, Unicode, UTF-8 encoded string that serves as the version ID of the backup plan |
vault_arn | The ARN of the vault |
vault_id | The name of the vault |
Known issues
During the developing of the module I found some issues reported to the The AWS provider:
Related backup plan selections must be deleted prior to backup plan deletion
$ terraform destroy
...
module.aws_backup_example.aws_iam_policy.ab_tag_policy: Destruction complete after 2s
module.aws_backup_example.aws_iam_role.ab_role: Destruction complete after 1s
Error: error deleting Backup Plan: InvalidRequestException: Related backup plan selections must be deleted prior to backup plan deletion
status code: 400, request id: 4a6637c8-2d46-4714-9929-4df3f694979b
When trying to destroy a plan, terraform complains about deleting the selections first, even though terraform tries to delete them in the right order:
This issue was reported as Backup Plan deletion fails randomly for the AWS Provider.
This happens because thee AWS provider tries to delete the plan without waiting for the selections destroyal confirmation.
Workaround:
I included and script in the examples that destroys the selections first and then destroys the plan:
cat terraform_destroy_aws_backup.sh
#!/bin/sh
targets=""
for i in `terraform state list | grep "selection"`; do targets="${targets} --target=${i}"; done
# Destroy selections
terraform destroy ${targets}
# Destroy all
terraform destroy
Error creating Backup Selection: IAM Role is not authorized to call tag:GetResources
Error: error creating Backup Selection: InvalidParameterValueException: IAM Role arn:aws:iam::111111111111:role/aws-backup-plan-complete-plan-role is not authorized to call tag:GetResources
status code: 400, request id: 07ab775d-8885-4240-bb99-41305df969e0
on .terraform/modules/aws_backup_example/selection.tf line 1, in resource "aws_backup_selection" "ab_selection":
1: resource "aws_backup_selection" "ab_selection" {
This issue was reported as aws_backup_selection.selection: error creating Backup Selection: InvalidParameterValueException for the AWS Provider.
I faced this when applying and destroying the same plan several times, for instance when I ws developing the module.
Workaround:
I couldn’t find any workaround for this. Just destroy all wait some time and apply again.
Leave a Comment