I would like to use a snapshot of an production RDS instance and load it into a processing cluster (use case: we need a stable copy of a database for long-running weekly process). I was able to create a cluster from the snapshot, then create the cluster instance the first time through. Next week, I found the existing processing cluster, modified it in TF using the latest snapshot of the production instance, and everything seemed to work. However, when I queried the database (either cluster or instance endpoints), it still had the contents from the original DB. The cluster_members attribute of the processing cluster indeed referenced the wrong (old) DB. All the other indications in state suggested that the recent snapshot had been successfully loaded.
How can I maintain an existing RDS cluster, and periodically replace the database instance using Terraform.
AWS, RDS Aurora Postgresql 9.6.x
TF 1.0.0