Conflict when creating Valkey cluster with Global replication

[Israphel]

Description

I'm trying to create a Valkey with cluster mode enabled and global replication.
I'm following the Redis global example which is the closest one I found here.

• ✋ I have searched the open/closed issues and my issue is not listed.

Versions

• Module version [Required]: 1.6.0

• Terraform version: 1.5.7

• Provider version(s): 6.0.0

Reproduction Code [Required]

Steps to reproduce the behavior:

# Elasticache (us-east-1)
module "elasticache-us-east-1" {
  source   = "terraform-aws-modules/elasticache/aws"
  version  = "1.6.0"
  for_each = local.config["global"]

  replication_group_id                    = each.key
  create_primary_global_replication_group = true
  cluster_mode_enabled                    = try(each.value.cluster_mode, false)
  cluster_id                              = try(each.value.cluster_mode, false) ? each.key : null
  engine                                  = try(each.value.engine, "valkey")
  engine_version                          = each.value.version
  node_type                               = each.value.node_type
  num_cache_clusters                      = !try(each.value.cluster_mode, false) ? each.value.nodes : null
  num_node_groups                         = try(each.value.cluster_mode, false) ? each.value.nodes : null
  replicas_per_node_group                 = try(each.value.cluster_mode, false) ? each.value.replicas_per_node : null
  az_mode                                 = "cross-az"
  maintenance_window                      = "sun:05:00-sun:09:00"
  auto_minor_version_upgrade              = false
  apply_immediately                       = true
  automatic_failover_enabled              = true
  create_security_group                   = false
  security_group_ids                      = ["sg-xxxx"]
  at_rest_encryption_enabled              = true
  transit_encryption_enabled              = true
  auth_token                              = "xxxx"
  log_delivery_configuration              = {}

  # Subnet Group
  subnet_group_name        = each.key
  subnet_group_description = "Subnet group for ${each.key}"
  subnet_ids               = ["xxxx", "xxxx", "xxxx", "xxxx"]

  # Parameter Group
  create_parameter_group      = true
  parameter_group_name        = "${each.key}-${replace("${try(each.value.engine, "valkey")}${replace(each.value.version, "/\\.[\\d]+$/", "")}", ".", "")}"
  parameter_group_description = "Customized parameter group for ${each.key}"
  parameter_group_family      = "${try(each.value.engine, "valkey")}${replace(each.value.version, "/\\.[\\d]+$/", "")}"
  parameters                  = try(each.value.parameters, [])

  tags = {
    Name        = each.key
    service     = try(each.value.service_name, each.key)
    environment = local.environment
    managed_by  = "terraform"
  }
}

# Elasticache (us-west-2)
module "elasticache-us-west-2" {
  source   = "terraform-aws-modules/elasticache/aws"
  version  = "1.6.0"
  for_each = local.config["global"]

  replication_group_id        = each.key
  global_replication_group_id = module.elasticache-us-east-1[each.key].global_replication_group_id
  create_security_group       = false
  security_group_ids          = ["sg-xxx"]
  log_delivery_configuration  = {}

  # Subnet Group
  subnet_group_name        = each.key
  subnet_group_description = "Subnet group for ${each.key}"
  subnet_ids               = ["xxxx", "xxxx", "xxxx", "xxxx"]

  tags = {
    Name        = each.key
    service     = try(each.value.service_name, each.key)
    environment = local.environment
    managed_by  = "terraform"
  }

  providers = {
    aws = aws.usw2
  }
}

global:
  maldo:
    engine: valkey
    version: "8.0"
    cluster_mode: true
    nodes: 2
    replicas_per_node: 1
    node_type: cache.m7g.large

Expected behavior

I should get two replication groups.

Actual behavior

The creation of the west replication group fails due to a conflict, even tho I'm not passing engine.

Terminal Output

│ Error: Conflicting configuration arguments
│
│   with module.elasticache-us-west-2["maldo"].aws_elasticache_replication_group.this[0],
│   on .terraform/modules/elasticache-us-west-2/main.tf line 95, in resource "aws_elasticache_replication_group" "this":
│   95:   global_replication_group_id = var.global_replication_group_id
│
│ "global_replication_group_id": conflicts with engine

[bryantbiggs]

you are setting the conflicting values - don't do that and it should be resolved

[Israphel]

you are setting the conflicting values - don't do that and it should be resolved

where? my us-west-2 call doesn't have engine.

# Elasticache (us-west-2)
module "elasticache-us-west-2" {
  source   = "terraform-aws-modules/elasticache/aws"
  version  = "1.6.0"
  for_each = local.config["global"]

  replication_group_id        = each.key
  global_replication_group_id = module.elasticache-us-east-1[each.key].global_replication_group_id
  create_security_group       = false
  security_group_ids          = ["sg-xxx"]
  log_delivery_configuration  = {}

  # Subnet Group
  subnet_group_name        = each.key
  subnet_group_description = "Subnet group for ${each.key}"
  subnet_ids               = ["xxxx", "xxxx", "xxxx", "xxxx"]

  tags = {
    Name        = each.key
    service     = try(each.value.service_name, each.key)
    environment = local.environment
    managed_by  = "terraform"
  }

  providers = {
    aws = aws.usw2
  }
}

[Israphel]

even after nullifying everything that conflicts, engine is set to redis instead of valkey

  # Fixes
  engine                      = null
  security_group_names        = null
  transit_encryption_enabled  = null
  at_rest_encryption_enabled  = null
  snapshot_arns               = null

  # module.elasticache-us-west-2["maldo"].aws_elasticache_replication_group.this[0] will be created
  + resource "aws_elasticache_replication_group" "this" {
      + apply_immediately              = (known after apply)
      + arn                            = (known after apply)
      + at_rest_encryption_enabled     = (known after apply)
      + auto_minor_version_upgrade     = (known after apply)
      + automatic_failover_enabled     = false
      + cluster_enabled                = (known after apply)
      + cluster_mode                   = (known after apply)
      + configuration_endpoint_address = (known after apply)
      + data_tiering_enabled           = (known after apply)
      + description                    = "Replication group"
      + engine                         = "redis"
      + engine_version                 = (known after apply)
      + engine_version_actual          = (known after apply)
      + global_replication_group_id    = "ldgnf-maldo"
      + id                             = (known after apply)
      + ip_discovery                   = (known after apply)
      + maintenance_window             = (known after apply)
      + member_clusters                = (known after apply)
      + multi_az_enabled               = false
      + network_type                   = (known after apply)
      + node_type                      = (known after apply)
      + num_cache_clusters             = (known after apply)
      + num_node_groups                = (known after apply)
      + parameter_group_name           = (known after apply)
      + port                           = 6379
      + preferred_cache_cluster_azs    = []
      + primary_endpoint_address       = (known after apply)
      + reader_endpoint_address        = (known after apply)
      + region                         = "us-west-2"
      + replicas_per_node_group        = (known after apply)
      + replication_group_id           = "maldo"
      + security_group_ids             = [
          + "sg-0f40b50581797dc68",
        ]
      + security_group_names           = (known after apply)
      + snapshot_window                = (known after apply)
      + subnet_group_name              = "maldo"
      + tags                           = {
          + "Name"                  = "maldo"
          + "environment"           = "stage"
          + "managed_by"            = "terraform"
          + "service"               = "maldo"
          + "terraform-aws-modules" = "elasticache"
        }
      + tags_all                       = {
          + "Name"                  = "maldo"
          + "environment"           = "stage"
          + "managed_by"            = "terraform"
          + "service"               = "maldo"
          + "terraform-aws-modules" = "elasticache"
        }
      + transit_encryption_enabled     = (known after apply)
      + transit_encryption_mode        = (known after apply)
    }

so I don't think this is going to work even if I set everything right.

you may be affected by this bug: hashicorp/terraform-provider-aws#41713