provider/aws: aws_elasticache_cluster normalizes name to lowercase

[thrashr888]

This is a similar issue to #3120. That patch fixes the aws_elasticache_subnet_group while this patch fixes the aws_elasticache_cluster. It's nearly the same change.

AWS forces aws_elasticache_cluster.cluster_id to be lower case. Each time terraform apply is run it cannot find the matching case cluster_id in it's state and tries to create a new one. This patch forces the cluster_id in Terraform to be lower case so the state can keep track of the correct cluster_id.

The cluster_id string in my project is generated with a variable with mixed case values, so I couldn't just make it lower case as a work-around.

I've tweaked the TestAccAWSElasticacheCluster_vpc acceptance test to include a mixed case version of the cluster_id. The acceptance tests pass when running make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSElasticacheCluster_vpc'. The regular tests pass as well.

[apparentlymart]

👍 LGTM!

[apparentlymart]

We had an apparent recurrence of this issue today:

-/+ aws_elasticache_cluster.web
    apply_immediately:      "" => "<computed>"
    cache_nodes.#:          "1" => "<computed>"
    cluster_id:             "omt-jsmith-dev" => "omt-jsmith-DEV" (forces new resource)
    configuration_endpoint: "omt-jsmith-dev.7dilrw.cfg.usw1.cache.amazonaws.com:11211" => "<computed>"
    engine:                 "memcached" => "memcached"
    engine_version:         "1.4.14" => "1.4.14"
    maintenance_window:     "tue:06:00-tue:07:00" => "<computed>"
    node_type:              "cache.m3.large" => "cache.m3.large"
    num_cache_nodes:        "1" => "1"
    parameter_group_name:   "default.memcached1.4" => "default.memcached1.4"
    port:                   "11211" => "11211"
    security_group_ids.#:   "0" => "<computed>"
    security_group_names.#: "0" => "<computed>"
    subnet_group_name:      "usw1-dev" => "usw1-dev"


Plan: 1 to add, 0 to change, 1 to destroy.

The config sets cluster_id to be omt-${var.env_name} where the environment name is set by an external system to jsmith-DEV. As we expect, the ElastiCache API normalized that to omt-jsmith-dev, but it looks like we didn't correctly apply the StateFunc to the config when making the diff, so Terraform wants to switch it back to uppercase again.

Just making a note of this for the moment so I don't lose this context. I'll try to circle back and look more closely at this in a little while.

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.