Terraform: Feature request: Support for ElastiCache Redis cluster mode

[bsiegel]

Amazon recently added support for Redis 3.2 and with it, support for the native Redis Cluster capability, allowing you to create a Redis ElastiCache cluster with 1-15 shards and 0-5 replicas per shard. The AWS documentation now makes the distinction between Memcache, Redis (cluster mode disabled), and Redis (cluster mode enabled). I would enjoy being able to create such a cluster via Terraform.

The announcement is here: https://aws.amazon.com/blogs/aws/amazon-elasticache-for-redis-update-sharded-clusters-engine-improvements-and-more/

Further reading here:
https://docs.aws.amazon.com/AmazonElastiCache/latest/UserGuide/Replication.Redis-RedisCluster.html
https://docs.aws.amazon.com/AmazonElastiCache/latest/UserGuide/SelectEngine.Uses.html

[jaygorrell]

Is this still being worked on? Would love to see this feature available!

[vamsirede]

Is this feature "ElastiCache Redis cluster mode" available??

[JarenGlover]

+1

[carhartl]

According to the cli reference* one is able to control the cluster mode via --cache-parameter-group, it seems this is already supported?

default.redis3.2.cluster.on cluster mode enabled
default.redis3.2 cluster mode disabled

=>

resource "aws_elasticache_cluster" "bar" {
    parameter_group_name = "default.redis3.2.cluster.on"
}

*http://docs.aws.amazon.com/AmazonElastiCache/latest/UserGuide/Clusters.Create.CLI.html#Clusters.Create.CLI.Redis

[mblakele]

According to the cli reference* one is able to control the cluster mode via --cache-parameter-group, it seems this is already supported?

Not quite, but see https://github.com/hashicorp/terraform/pull/9892 for a candidate implementation.

Would be great to see this as part of terraform 0.9.

[rafaelfc-olx]

Any news on this? Any intention of enabling this into terraform 0.8 as well?

[deepybee]

It's taken me while to poke around but this is actually already supported in 0.8.8 (possibly earlier):

resource "aws_elasticache_replication_group" "foo-replication-group" {
  replication_group_id          = "foo"
  replication_group_description = "Foo replication group"
  node_type                     = "cache.m1.small"
  number_cache_clusters         = 2
  port                          = 6379
  parameter_group_name          = "default.redis3.2.cluster.on"
  availability_zones            = ["eu-west-1a", "eu-west-1b"]
  automatic_failover_enabled    = true
  security_group_ids            = ["${aws_security_group.foo_sg.id}"]
  subnet_group_name             = "${aws_elasticache_subnet_group.foo-subnets.name}"
}

EDIT: the key point is, I think, the confusion stems from attempts to configure clustered mode with an aws_elasticache_cluster resource when an aws_elasticache_replication_group resource should be used instead.

[randahlem]

@deepybee Do you have any ideas on how to scale up the number_cache_clusters minimizing the downtime of destroying/creating the new resource? Thanks in advance!

[randahlem]

It seems to be still missing the real clustering part. According to this 9892

[deepybee]

Yes @randahlem, it looks like that's a destructive change at present. However, I've manually spun a cluster mode redis cluster in AWS and the modification options are pretty limited. The button to add or remove nodes to shard across is greyed out, so this may be a limitation of the Elasticache API rather than Terraform.

[stack72]

@bsiegel I believe this has been added now - please can you check this when 0.9.6 comes out?

[msolimans]

I deployed it in a clustered mode using the following:

resource "aws_elasticache_replication_group" "foo" {
...
...
cluster_mode {
    num_node_groups = 2
    replicas_per_node_group = 1
  }

}

[hashibot[bot]]

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.