aws_db_instance restore from snapshot intermittent state collision.

[scott-kausler]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform CLI and Terraform AWS Provider Version

Terraform CLI Versionv0.12.20
Terraform AWS Provider Version 2.69.0

Affected Resource(s)

aws_db_instance

Terraform Configuration Files

resource "aws_db_instance" "this" {
...
  snapshot_identifier       = var.snapshot_arn

  parameter_group_name   = var.parameter_group_name
  backup_retention_period   = 7

  # When restoring from snapshot, terraform tries to set the password, but the provider could error out because of this.
  username = var.username
  password =  var.password
}

Debug Output

[2020-08-20T03:42:14.996Z] module.rds.aws_db_instance.this: Still creating... [9m50s elapsed]
[2020-08-20T03:42:21.517Z] module.rds.aws_db_instance.this: Creation complete after 9m58s [id=db-instance]
[2020-08-20T03:42:24.891Z] module.rds.aws_db_instance.this: Still creating... [10m0s elapsed]
[2020-08-20T03:42:34.826Z] module.rds.aws_db_instance.this: Still creating... [10m10s elapsed]
[2020-08-20T03:42:44.714Z] module.rds.aws_db_instance.this: Still creating... [10m20s elapsed]
[2020-08-20T03:42:55.028Z] module.rds.aws_db_instance.this: Still creating... [10m30s elapsed]
[2020-08-20T03:43:04.850Z] module.rds.aws_db_instance.this: Still creating... [10m40s elapsed]
[2020-08-20T03:43:05.775Z] module.rds.aws_db_instance.this: Refreshing state... [id=db-instance]
[2020-08-20T03:43:09.292Z] module.rds.aws_db_instance.this: Modifying... [id=db-instance]
[2020-08-20T03:43:09.292Z]
[2020-08-20T03:43:09.292Z] Error: Error modifying DB Instance db-instance: InvalidDBInstanceState: Database instance is not in available state.
[2020-08-20T03:43:09.292Z] status code: 400, request id: 7d9c0234-e364-49e7-9eb6-e8d7886a8d0e
[2020-08-20T03:43:09.292Z]
[2020-08-20T03:43:09.292Z] on .terraform/modules/rds/main.tf line 1, in resource "aws_db_instance" "this":
[2020-08-20T03:43:09.292Z] 1: resource "aws_db_instance" "this" {
[2020-08-20T03:43:09.292Z]

Expected Behavior

The db instance should have been restored successfully

Actual Behavior

Because the password was set and the backup_retention_period differed from the snapshot, the aws provider attempted to modify both, presumably without the correct wait between modifying.

Steps to Reproduce

1. terraform apply

Important Factoids

• When modifying the DB instance, the status may remain "Available" before moving to "Modifying" and then back to "Available" once the operation is complete. The provider should expect that behavior.
• Presumably, this could be reproduced with any parameter that differs from the snapshot, such as allocated_storage, although it was encountered when only modifying parameter_group_name, backup_retention_period, and password.

[ljluestc]

# Define a data source to fetch snapshot information
data "aws_db_snapshot" "example" {
  snapshot_identifier = var.snapshot_arn
}

# Wait for the DB instance to become available
resource "null_resource" "wait_for_db" {
  provisioner "local-exec" {
    command = "sleep 300"  # Adjust the sleep duration as needed
  }
  depends_on = [aws_db_instance.this]
}

# Create the DB instance
resource "aws_db_instance" "this" {
  // Other DB instance attributes

  # Set the backup retention period based on the snapshot
  backup_retention_period = data.aws_db_snapshot.example.allocated_storage

  # Set the password based on the snapshot
  password = data.aws_db_snapshot.example.encrypted
}

# Optionally, you can add a local-exec provisioner to wait for the DB instance
# to become available before proceeding with other tasks.

use the aws_db_snapshot data source to fetch information about the snapshot, such as the allocated storage and encrypted attributes.
add a null_resource with a local-exec provisioner to wait for a particular duration (e.g., 300 seconds) to allow the DB instance to become fully available before proceeding with other tasks. You can adjust the sleep duration as needed.