Upgrading Redis Engine/Family to a new version fails

[gavllew]

What is the current behavior?

Changing the Redis engine_version/family of an existing cluster fails, as it tries to delete the Parameter Group while it is still used by the existing cluster:

│ Error: deleting ElastiCache Parameter Group (pfdev-whiteboard-redis-c499): InvalidCacheParameterGroupState: One or more cache clusters are still members of this parameter group pfdev-whiteboard-redis-c499, so the group cannot be deleted.

PR #48 removed the create_before_destroy lifecycle setting on this resource.

If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem.

1. Deploy a Redis cluster using this module
2. Update the engine_version/family of the cluster and attempt a Terraform Apply

What is the expected behavior?

A graceful upgrade to a new Redis major version is possible.

Software versions?

umotif-public/elasticache-redis/aws v3.5.0