Fixed the documentation for ElastiCache Replication Group address

[Ninir]

Fixes #375

As exposed in the code, primary_endpoint_address is set when there are node groups, so when in a cluster mode.
configuration_endpoint_address is the exact opposite.

[radeksimko]

👍 Thanks.

[phunehehe]

I'm not sure what I'm missing but I'm seeing the opposite: configuration endpoint only exists if cluster mode is enabled, and primary endpoint only exists if cluster mode is disabled...

[radeksimko]

@phunehehe Do you mind sharing a config which leads you to that conclusion?

I used the following (practically clustered Redis with 2 nodes):

resource "aws_vpc" "network" {
  cidr_block = "10.10.0.0/16"
}

resource "aws_subnet" "main" {
  cidr_block = "10.10.0.0/24"
  availability_zone = "eu-west-2a"
  vpc_id = "${aws_vpc.network.id}"
}

resource "aws_subnet" "secondary" {
  cidr_block = "10.10.5.0/24"
  availability_zone = "eu-west-2b"
  vpc_id = "${aws_vpc.network.id}"
}

resource "aws_elasticache_subnet_group" "main" {
  name = "radek-test"
  subnet_ids = ["${aws_subnet.main.id}", "${aws_subnet.secondary.id}"]
}

resource "aws_elasticache_replication_group" "autoscaling_event_update_route53" {
  replication_group_id          = "asgevnt-route53"
  replication_group_description = "Autoscaling events updating Route53"

  engine_version                = "3.2.4"
  node_type                     = "cache.m4.large"
  number_cache_clusters         = "2"
  port                          = 6379

  subnet_group_name             = "${aws_elasticache_subnet_group.main.name}"
  parameter_group_name          = "default.redis3.2"

  automatic_failover_enabled    = "true"
  availability_zones            = ["eu-west-2a", "eu-west-2b"]
  apply_immediately             = "true"
  maintenance_window            = "sun:01:00-sun:02:00"

  tags {
    Name                        = "autoscaling_event_update_route53"
    environment                 = "core"
    service                     = "main"
  }
}

and it does expose the primary_endpoint_address:

$ echo 'aws_elasticache_replication_group.autoscaling_event_update_route53.primary_endpoint_address' | terraform console
asgevnt-route53.g4wffw.ng.0001.euw2.cache.amazonaws.com

so I believe @Ninir is right.

[phunehehe]

So I created a setup with various configurations https://gist.github.com/phunehehe/9657cb33d7ba0ab48a113e9c3bc54475 and here's the output:

single	primary_endpoint_address	yes
single	configuration_endpoint_address	no
cluster	primary_endpoint_address	yes
cluster	configuration_endpoint_address	no
native_cluster	primary_endpoint_address	yes
native_cluster	configuration_endpoint_address	no
native_single	primary_endpoint_address	yes
native_single	configuration_endpoint_address	no
sharding	primary_endpoint_address	no
sharding	configuration_endpoint_address	yes

It would seem like configuration_endpoint_address exists when sharding is enabled, and primary_endpoint_address exists otherwise. When sharding is enabled, AWS console displays "Clustered Redis" in the "Mode" column.

[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 feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thanks!