Merge pull request #1075 from iht/bigtable_replication

Add cluster replicas to Bigtable module.
This commit is contained in:
Israel Herraiz 2022-12-30 11:39:37 +01:00 committed by GitHub
commit 0380c1004b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
4 changed files with 138 additions and 86 deletions

View File

@ -5,7 +5,6 @@ This module allows managing a single BigTable instance, including access configu
## TODO
- [ ] support bigtable_app_profile
- [ ] support cluster replicas
- [ ] support IAM for tables
## Examples
@ -18,8 +17,11 @@ module "bigtable-instance" {
source = "./fabric/modules/bigtable-instance"
project_id = "my-project"
name = "instance"
cluster_id = "instance"
zone = "europe-west1-b"
clusters = {
my-cluster = {
zone = "europe-west1-b"
}
}
tables = {
test1 = {},
test2 = {
@ -41,8 +43,11 @@ module "bigtable-instance" {
source = "./fabric/modules/bigtable-instance"
project_id = "my-project"
name = "instance"
cluster_id = "instance"
zone = "europe-west1-b"
clusters = {
my-cluster = {
zone = "europe-west1-b"
}
}
tables = {
test1 = {},
test2 = {
@ -63,6 +68,29 @@ module "bigtable-instance" {
# tftest modules=1 resources=4
```
### Instance with replication enabled
```hcl
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
```
### Instance with garbage collection policy
```hcl
@ -71,8 +99,11 @@ module "bigtable-instance" {
source = "./fabric/modules/bigtable-instance"
project_id = "my-project"
name = "instance"
cluster_id = "instance"
zone = "europe-west1-b"
clusters = {
my-cluster = {
zone = "europe-west1-b"
}
}
tables = {
test1 = {
column_families = {
@ -102,8 +133,11 @@ module "bigtable-instance" {
source = "./fabric/modules/bigtable-instance"
project_id = "my-project"
name = "instance"
cluster_id = "instance"
zone = "europe-west1-b"
clusters = {
my-cluster = {
zone = "europe-west1-b"
}
}
default_gc_policy = {
deletion_policy = "ABANDON"
max_age = "18h"
@ -131,9 +165,12 @@ module "bigtable-instance" {
source = "./fabric/modules/bigtable-instance"
project_id = "my-project"
name = "instance"
cluster_id = "instance"
zone = "europe-west1-b"
num_nodes = 5
clusters = {
my-cluster = {
zone = "europe-west1-b"
num_nodes = 5
}
}
}
# tftest modules=1 resources=1
```
@ -148,13 +185,18 @@ module "bigtable-instance" {
source = "./fabric/modules/bigtable-instance"
project_id = "my-project"
name = "instance"
cluster_id = "instance"
zone = "europe-southwest1-b"
autoscaling_config = {
min_nodes = 3
max_nodes = 7
cpu_target = 70
clusters = {
my-cluster = {
zone = "europe-southwest1-b"
autoscaling = {
min_nodes = 3
max_nodes = 7
cpu_target = 70
}
}
}
}
# tftest modules=1 resources=1
```
@ -164,17 +206,20 @@ module "bigtable-instance" {
```hcl
module "bigtable-instance" {
source = "./fabric/modules/bigtable-instance"
project_id = "my-project"
name = "instance"
cluster_id = "instance"
zone = "europe-southwest1-a"
storage_type = "SSD"
autoscaling_config = {
min_nodes = 3
max_nodes = 7
cpu_target = 70
storage_target = 4096
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
@ -185,19 +230,16 @@ module "bigtable-instance" {
| name | description | type | required | default |
|---|---|:---:|:---:|:---:|
| [name](variables.tf#L68) | The name of the Cloud Bigtable instance. | <code>string</code> | ✓ | |
| [project_id](variables.tf#L79) | Id of the project where datasets will be created. | <code>string</code> | ✓ | |
| [zone](variables.tf#L109) | The zone to create the Cloud Bigtable cluster in. | <code>string</code> | ✓ | |
| [autoscaling_config](variables.tf#L17) | Settings for autoscaling of the instance. If you set this variable, the variable num_nodes is ignored. | <code title="object&#40;&#123;&#10; min_nodes &#61; number&#10; max_nodes &#61; number&#10; cpu_target &#61; number&#10; storage_target &#61; optional&#40;number, null&#41;&#10;&#125;&#41;">object&#40;&#123;&#8230;&#125;&#41;</code> | | <code>null</code> |
| [cluster_id](variables.tf#L28) | The ID of the Cloud Bigtable cluster. | <code>string</code> | | <code>&#34;europe-west1&#34;</code> |
| [default_gc_policy](variables.tf#L34) | 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. | <code title="object&#40;&#123;&#10; deletion_policy &#61; optional&#40;string&#41;&#10; gc_rules &#61; optional&#40;string&#41;&#10; mode &#61; optional&#40;string&#41;&#10; max_age &#61; optional&#40;string&#41;&#10; max_version &#61; optional&#40;string&#41;&#10;&#125;&#41;">object&#40;&#123;&#8230;&#125;&#41;</code> | | <code>null</code> |
| [deletion_protection](variables.tf#L46) | 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. | <code></code> | | <code>true</code> |
| [display_name](variables.tf#L51) | The human-readable display name of the Bigtable instance. | <code></code> | | <code>null</code> |
| [iam](variables.tf#L56) | IAM bindings for topic in {ROLE => [MEMBERS]} format. | <code>map&#40;list&#40;string&#41;&#41;</code> | | <code>&#123;&#125;</code> |
| [instance_type](variables.tf#L62) | (deprecated) The instance type to create. One of 'DEVELOPMENT' or 'PRODUCTION'. | <code>string</code> | | <code>null</code> |
| [num_nodes](variables.tf#L73) | The number of nodes in your Cloud Bigtable cluster. This value is ignored if you are using autoscaling. | <code>number</code> | | <code>1</code> |
| [storage_type](variables.tf#L84) | The storage type to use. | <code>string</code> | | <code>&#34;SSD&#34;</code> |
| [tables](variables.tf#L90) | Tables to be created in the BigTable instance. | <code title="map&#40;object&#40;&#123;&#10; split_keys &#61; optional&#40;list&#40;string&#41;, &#91;&#93;&#41;&#10; column_families &#61; optional&#40;map&#40;object&#40;&#10; &#123;&#10; gc_policy &#61; optional&#40;object&#40;&#123;&#10; deletion_policy &#61; optional&#40;string&#41;&#10; gc_rules &#61; optional&#40;string&#41;&#10; mode &#61; optional&#40;string&#41;&#10; max_age &#61; optional&#40;string&#41;&#10; max_version &#61; optional&#40;string&#41;&#10; &#125;&#41;, null&#41;&#10; &#125;&#41;&#41;, &#123;&#125;&#41;&#10;&#125;&#41;&#41;">map&#40;object&#40;&#123;&#8230;&#125;&#41;&#41;</code> | | <code>&#123;&#125;</code> |
| [clusters](variables.tf#L17) | 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. | <code title="map&#40;object&#40;&#123;&#10; zone &#61; optional&#40;string&#41;&#10; storage_type &#61; optional&#40;string&#41;&#10; num_nodes &#61; optional&#40;number&#41;&#10; autoscaling &#61; optional&#40;object&#40;&#123;&#10; min_nodes &#61; number&#10; max_nodes &#61; number&#10; cpu_target &#61; number&#10; storage_target &#61; optional&#40;number&#41;&#10; &#125;&#41;&#41;&#10;&#125;&#41;&#41;">map&#40;object&#40;&#123;&#8230;&#125;&#41;&#41;</code> | ✓ | |
| [name](variables.tf#L78) | The name of the Cloud Bigtable instance. | <code>string</code> | ✓ | |
| [project_id](variables.tf#L83) | Id of the project where datasets will be created. | <code>string</code> | ✓ | |
| [default_autoscaling](variables.tf#L33) | Default settings for autoscaling of clusters. This will be the default autoscaling for any cluster not specifying any autoscaling details. | <code title="object&#40;&#123;&#10; min_nodes &#61; number&#10; max_nodes &#61; number&#10; cpu_target &#61; number&#10; storage_target &#61; optional&#40;number&#41;&#10;&#125;&#41;">object&#40;&#123;&#8230;&#125;&#41;</code> | | <code>null</code> |
| [default_gc_policy](variables.tf#L44) | 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. | <code title="object&#40;&#123;&#10; deletion_policy &#61; optional&#40;string&#41;&#10; gc_rules &#61; optional&#40;string&#41;&#10; mode &#61; optional&#40;string&#41;&#10; max_age &#61; optional&#40;string&#41;&#10; max_version &#61; optional&#40;string&#41;&#10;&#125;&#41;">object&#40;&#123;&#8230;&#125;&#41;</code> | | <code>null</code> |
| [deletion_protection](variables.tf#L56) | 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. | <code></code> | | <code>true</code> |
| [display_name](variables.tf#L61) | The human-readable display name of the Bigtable instance. | <code></code> | | <code>null</code> |
| [iam](variables.tf#L66) | IAM bindings for topic in {ROLE => [MEMBERS]} format. | <code>map&#40;list&#40;string&#41;&#41;</code> | | <code>&#123;&#125;</code> |
| [instance_type](variables.tf#L72) | (deprecated) The instance type to create. One of 'DEVELOPMENT' or 'PRODUCTION'. | <code>string</code> | | <code>null</code> |
| [tables](variables.tf#L88) | Tables to be created in the BigTable instance. | <code title="map&#40;object&#40;&#123;&#10; split_keys &#61; optional&#40;list&#40;string&#41;, &#91;&#93;&#41;&#10; column_families &#61; optional&#40;map&#40;object&#40;&#10; &#123;&#10; gc_policy &#61; optional&#40;object&#40;&#123;&#10; deletion_policy &#61; optional&#40;string&#41;&#10; gc_rules &#61; optional&#40;string&#41;&#10; mode &#61; optional&#40;string&#41;&#10; max_age &#61; optional&#40;string&#41;&#10; max_version &#61; optional&#40;string&#41;&#10; &#125;&#41;, null&#41;&#10; &#125;&#41;&#41;, &#123;&#125;&#41;&#10;&#125;&#41;&#41;">map&#40;object&#40;&#123;&#8230;&#125;&#41;&#41;</code> | | <code>&#123;&#125;</code> |
## Outputs

View File

@ -15,7 +15,6 @@
*/
locals {
num_nodes = var.autoscaling_config == null ? var.num_nodes : null
gc_pairs = flatten([
for table, table_obj in var.tables : [
for cf, cf_obj in table_obj.column_families : {
@ -25,30 +24,44 @@ locals {
}
]
])
clusters_autoscaling = {
for cluster_id, cluster in var.clusters : cluster_id => {
zone = cluster.zone
storage_type = cluster.storage_type
num_nodes = cluster.autoscaling == null && var.default_autoscaling == null ? cluster.num_nodes : null
autoscaling = cluster.autoscaling == null ? var.default_autoscaling : cluster.autoscaling
}
}
}
resource "google_bigtable_instance" "default" {
project = var.project_id
name = var.name
cluster {
cluster_id = var.cluster_id
zone = var.zone
storage_type = var.storage_type
num_nodes = local.num_nodes
dynamic "autoscaling_config" {
for_each = var.autoscaling_config == null ? [] : [""]
content {
min_nodes = var.autoscaling_config.min_nodes
max_nodes = var.autoscaling_config.max_nodes
cpu_target = var.autoscaling_config.cpu_target
storage_target = var.autoscaling_config.storage_target
instance_type = var.instance_type
display_name = var.display_name == null ? var.display_name : var.name
deletion_protection = var.deletion_protection
dynamic "cluster" {
for_each = local.clusters_autoscaling
content {
cluster_id = cluster.key
zone = cluster.value.zone
storage_type = cluster.value.storage_type
num_nodes = cluster.value.num_nodes
dynamic "autoscaling_config" {
for_each = cluster.value.autoscaling == null ? [] : [""]
content {
min_nodes = cluster.value.autoscaling.min_nodes
max_nodes = cluster.value.autoscaling.max_nodes
cpu_target = cluster.value.autoscaling.cpu_target
storage_target = cluster.value.autoscaling.storage_target
}
}
}
}
instance_type = var.instance_type
display_name = var.display_name == null ? var.display_name : var.name
deletion_protection = var.deletion_protection
}
resource "google_bigtable_instance_iam_binding" "default" {

View File

@ -14,23 +14,33 @@
* limitations under the License.
*/
variable "autoscaling_config" {
description = "Settings for autoscaling of the instance. If you set this variable, the variable num_nodes is ignored."
variable "clusters" {
description = "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."
nullable = false
type = map(object({
zone = optional(string)
storage_type = optional(string)
num_nodes = optional(number)
autoscaling = optional(object({
min_nodes = number
max_nodes = number
cpu_target = number
storage_target = optional(number)
}))
}))
}
variable "default_autoscaling" {
description = "Default settings for autoscaling of clusters. This will be the default autoscaling for any cluster not specifying any autoscaling details."
type = object({
min_nodes = number
max_nodes = number
cpu_target = number
storage_target = optional(number, null)
storage_target = optional(number)
})
default = null
}
variable "cluster_id" {
description = "The ID of the Cloud Bigtable cluster."
type = string
default = "europe-west1"
}
variable "default_gc_policy" {
description = "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."
type = object({
@ -70,23 +80,11 @@ variable "name" {
type = string
}
variable "num_nodes" {
description = "The number of nodes in your Cloud Bigtable cluster. This value is ignored if you are using autoscaling."
type = number
default = 1
}
variable "project_id" {
description = "Id of the project where datasets will be created."
type = string
}
variable "storage_type" {
description = "The storage type to use."
type = string
default = "SSD"
}
variable "tables" {
description = "Tables to be created in the BigTable instance."
nullable = false
@ -105,8 +103,3 @@ variable "tables" {
}))
default = {}
}
variable "zone" {
description = "The zone to create the Cloud Bigtable cluster in."
type = string
}

View File

@ -28,5 +28,9 @@ module "test" {
}
}
zone = var.zone
clusters = {
test = {
zone = var.zone
}
}
}