Skip to content

Latest commit

 

History

History
252 lines (216 loc) · 8.39 KB

File metadata and controls

252 lines (216 loc) · 8.39 KB

Google Cloud BigTable Module

This module allows managing a single BigTable instance, including access configuration and tables.

TODO

  • support bigtable_app_profile
  • support IAM for tables

Examples

Instance with access configuration

module "bigtable-instance" {
  source     = "./fabric/modules/bigtable-instance"
  project_id = "my-project"
  name       = "instance"
  clusters = {
    my-cluster = {
      zone = "europe-west1-b"
    }
  }
  tables = {
    test1 = {},
    test2 = {
      split_keys = ["a", "b", "c"]
    }
  }
  iam = {
    "roles/bigtable.user" = ["user:[email protected]"]
  }
}
# tftest modules=1 resources=4 inventory=simple.yaml

Instance with tables and column families

module "bigtable-instance" {
  source     = "./fabric/modules/bigtable-instance"
  project_id = "my-project"
  name       = "instance"
  clusters = {
    my-cluster = {
      zone = "europe-west1-b"
    }
  }
  tables = {
    test1 = {},
    test2 = {
      split_keys = ["a", "b", "c"]
      column_families = {
        cf1 = {}
        cf2 = {}
        cf3 = {}
      }
    }
    test3 = {
      column_families = {
        cf1 = {}
      }
    }
  }
}
# tftest modules=1 resources=4 inventory=columns.yaml

Instance with replication enabled

module "bigtable-instance" {
  source     = "./fabric/modules/bigtable-instance"
  project_id = "my-project"
  name       = "instance"
  clusters = {
    first-cluster = {
      zone = "europe-west1-b"
    }
    second-cluster = {
      zone = "europe-southwest1-a"
    }
    third-cluster = {
      zone = "us-central1-b"
    }
  }
}
# tftest modules=1 resources=1 inventory=replication.yaml

Instance with garbage collection policy

module "bigtable-instance" {
  source     = "./fabric/modules/bigtable-instance"
  project_id = "my-project"
  name       = "instance"
  clusters = {
    my-cluster = {
      zone = "europe-west1-b"
    }
  }
  tables = {
    test1 = {
      column_families = {
        cf1 = {
          gc_policy = {
            deletion_policy = "ABANDON"
            max_age         = "18h"
          }
        }
        cf2 = {}
      }
    }
  }
}
# tftest modules=1 resources=3 inventory=gc.yaml

Instance with default garbage collection policy

The default garbage collection policy is applied to any column family that does not specify a gc_policy. If a column family specifies a gc_policy, the default garbage collection policy is ignored for that column family.

module "bigtable-instance" {
  source     = "./fabric/modules/bigtable-instance"
  project_id = "my-project"
  name       = "instance"
  clusters = {
    my-cluster = {
      zone = "europe-west1-b"
    }
  }
  default_gc_policy = {
    deletion_policy = "ABANDON"
    max_age         = "18h"
    max_version     = 7
  }
  tables = {
    test1 = {
      column_families = {
        cf1 = {}
        cf2 = {}
      }
    }
  }
}
# tftest modules=1 resources=4

Instance with static number of nodes

If you are not using autoscaling settings, you must set a specific number of nodes with the variable num_nodes.

module "bigtable-instance" {
  source     = "./fabric/modules/bigtable-instance"
  project_id = "my-project"
  name       = "instance"
  clusters = {
    my-cluster = {
      zone      = "europe-west1-b"
      num_nodes = 5
    }
  }
}
# tftest modules=1 resources=1 inventory=static.yaml

Instance with autoscaling (based on CPU only)

If you use autoscaling, you should not set the variable num_nodes.

module "bigtable-instance" {
  source     = "./fabric/modules/bigtable-instance"
  project_id = "my-project"
  name       = "instance"
  clusters = {
    my-cluster = {
      zone = "europe-southwest1-b"
      autoscaling = {
        min_nodes  = 3
        max_nodes  = 7
        cpu_target = 70
      }
    }
  }


}
# tftest modules=1 resources=1 inventory=autoscaling1.yaml

Instance with autoscaling (based on CPU and/or storage)

module "bigtable-instance" {
  source     = "./fabric/modules/bigtable-instance"
  project_id = "my-project"
  name       = "instance"
  clusters = {
    my-cluster = {
      zone         = "europe-southwest1-a"
      storage_type = "SSD"
      autoscaling = {
        min_nodes      = 3
        max_nodes      = 7
        cpu_target     = 70
        storage_target = 4096
      }
    }
  }
}
# tftest modules=1 resources=1 inventory=autoscaling2.yaml

Variables

name description type required default
clusters Clusters to be created in the BigTable instance. Set more than one cluster to enable replication. If you set autoscaling, num_nodes will be ignored. map(object({…}))
name The name of the Cloud Bigtable instance. string
project_id Id of the project where datasets will be created. string
default_autoscaling Default settings for autoscaling of clusters. This will be the default autoscaling for any cluster not specifying any autoscaling details. object({…}) null
default_gc_policy Default garbage collection policy, to be applied to all column families and all tables. Can be override in the tables variable for specific column families. object({…}) null
deletion_protection Whether or not to allow Terraform to destroy the instance. Unless this field is set to false in Terraform state, a terraform destroy or terraform apply that would delete the instance will fail. bool true
display_name The human-readable display name of the Bigtable instance. string null
encryption_key The KMS key id to used for encryption of the Bigtable instance. string null
iam IAM bindings for topic in {ROLE => [MEMBERS]} format. map(list(string)) {}
labels Labels to be attached to the instance. map(string) {}
tables Tables to be created in the BigTable instance. map(object({…})) {}

Outputs

name description sensitive
id Fully qualified instance id.
instance BigTable instance.
table_ids Map of fully qualified table ids keyed by table name.
tables Table resources.