Intermittent MissingRequestBodyError while creating aws_s3_bucket_replication_configuration

[nvnivs]

Description

Occasionally, when creating a new S3 bucket with both Replication and Lifecycle configurations the apply fails with the following error:

│ Error: creating S3 Bucket (origin-s3-bucket-dear-pipefish) Replication Configuration: operation error S3: PutBucketReplication, https response error StatusCode: 400, RequestID: ---REDACTED---, HostID: ---REDACTED---, api error MissingRequestBodyError: Request Body is empty
│ 
│   with module.bucket_source.aws_s3_bucket_replication_configuration.this[0],
│   on ../../main.tf line 413, in resource "aws_s3_bucket_replication_configuration" "this":
│  413: resource "aws_s3_bucket_replication_configuration" "this" {

The root cause seems to be when the aws_s3_bucket_replication_configuration and the aws_s3_bucket_lifecycle_configuration resources creating a race condition when created in parallel.

Versions

• Module version [Required]:

• v5.7.0 (latest)

• v4.3.0 (oldest version we tested with)

• Terraform version:

Terraform v1.7.3

• Provider version(s):

Providers required by configuration:
.
└── provider[registry.terraform.io/hashicorp/aws] >= 6.5.0

Reproduction Code [Required]

Steps to reproduce the behavior:

• Create a configuration with both replication_configuration and lifecycle_rule
• Init
• Apply

Expected behavior

The module to apply successfully with the same configuration every time

Actual behavior

The module module creation only succeeds if aws_s3_bucket_replication_configuration creation is completed before aws_s3_bucket_lifecycle_configuration starts.

Additional context

Have created an example replicating the issue in a fork of this repo.

Also, have added an explicit dependency to the aws_s3_bucket_lifecycle_configuration which prevents this issue: https://github.com/nvnivs/terraform-aws-s3-bucket/blob/c255cb1e82698aaaea6feb095698398ebb318148/main.tf#L391-L396