AWS provider used for S3 backend in terraform_remote_state is incorrectly instantiated

[ghost]

There is a really ugly authentication problem which has been languishing, unresolved and unassigned, for nearly 2 years over in the AWS provider repo which actually belongs over here.

hashicorp/terraform-provider-aws#275

The problem is not in the AWS provider. The problem is in the manner in which the terraform provider instantiates an AWS provider from a completely separate config within the resource declaration. That is done in a manner which ignores the aws profile that has been requested for authentication which, in turn, completely breaks terraform_remote_state for any user that isn't using the default provider or who needs to specify more than a single profile via the AWS_PROFILE environment var. It's a relatively trivial fix which has been fully documented by various sufferers of the bug. At this point, it should take only minutes for someone familiar with the provider build and test processes to correct the problem, which is really a critical bug for the vast majority of users who encounter it.

Can someone please move that bug over to here and/or just correct the problem?

Thanks.

[wjam]

Having experienced a similar issue to this to do with how the s3 backend behaves, the solution to this is to update the github.com/hashicorp/terraform/* dependencies to v0.11.8 once it has been released.