Fix Terraform Elasticache Cluster - CacheClusterAlreadyExists

Quick Answer

An ElastiCache cluster with the same ID already exists. Import it into Terraform, choose a unique cluster ID, or delete the orphaned cluster. ElastiCache IDs are unique per region per account.

The Error

Error: creating ElastiCache Cluster (redis-prod):
  CacheClusterAlreadyExists: CacheCluster redis-prod already exists.

What Causes This Error

1. Cluster created manually via Console or CLI, now Terraform tries to create one with the same name
2. Another Terraform workspace already manages a cluster with this name
3. Failed previous apply created the cluster but didn't record it in state
4. Recently deleted cluster — ElastiCache may take minutes to fully release the name

How to Fix It

Solution 1: Import the Existing Cluster

aws elasticache describe-cache-clusters --cache-cluster-id redis-prod
 
terraform import aws_elasticache_cluster.redis redis-prod

Solution 2: Use Unique Naming

resource "aws_elasticache_cluster" "redis" {
  cluster_id           = "${var.project}-${var.environment}-redis"
  engine               = "redis"
  node_type            = "cache.t3.micro"
  num_cache_nodes      = 1
  parameter_group_name = "default.redis7"
  port                 = 6379
  subnet_group_name    = aws_elasticache_subnet_group.redis.name
  security_group_ids   = [aws_security_group.redis.id]
}

Solution 3: Use Replication Group Instead

For production Redis, use a replication group (supports failover):

resource "aws_elasticache_replication_group" "redis" {
  replication_group_id = "${var.project}-${var.environment}"
  description          = "Redis for ${var.project}"
  node_type            = "cache.r6g.large"
  num_cache_clusters   = 2
  engine_version       = "7.0"
  port                 = 6379
  subnet_group_name    = aws_elasticache_subnet_group.redis.name
  security_group_ids   = [aws_security_group.redis.id]
 
  automatic_failover_enabled = true
  at_rest_encryption_enabled = true
  transit_encryption_enabled = true
}

Troubleshooting Checklist

1. ✅ Does the cluster exist? (aws elasticache describe-cache-clusters)
2. ✅ Should you import it or delete it?
3. ✅ Are you using environment-specific names?
4. ✅ Was a previous apply interrupted?

Prevention Tips

• Include environment in cluster IDs — myapp-prod-redis not redis
• Use replication groups for production — better availability and naming
• Tag with ManagedBy = "terraform" for easy identification
• Import existing clusters before creating new ones

Related Articles

• All Troubleshooting Articles
• Terraform Glossary

Conclusion

CacheClusterAlreadyExists means the name is taken. Import the existing cluster or use unique environment-prefixed names. For production Redis, prefer replication groups over standalone clusters.