shohei-ihaya / dotfiles Goto Github PK

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request.
• Please do not leave +1 or me too comments, they generate extra noise for issue followers and do not help prioritize the request.
• If you are interested in working on this issue or have submitted a pull request, please leave a comment.
• If an issue is assigned to the modular-magician user, it is either in the process of being autogenerated, or is planned to be autogenerated soon. If an issue is assigned to a user, that user is claiming responsibility for the issue. If an issue is assigned to hashibot, a community member has claimed the issue already.

Terraform Version

Terraform v1.1.7

Affected Resource(s)

• google_sql_database_instance

Terraform Configuration Files

data "google_project" "project" {}

resource "google_sql_database_instance" "master" {
  project          = data.google_project.project.project_id
  region           = "asia-northeast1"
  name             = "mts-issue-test-master-001"
  database_version = "MYSQL_5_7"

  settings {
    availability_type = "ZONAL"
    disk_autoresize   = true
    disk_size         = 10
    disk_type         = "PD_SSD"
    tier              = "db-f1-micro"

    activation_policy = "ALWAYS"
    pricing_plan      = "PER_USE"

    backup_configuration {
      binary_log_enabled = true
      enabled            = true
      location           = "asia"
      start_time         = "18:00"
    }


    database_flags {
      name  = "character_set_server"
      value = "utf8mb4"
    }
  }
}


resource "google_sql_database_instance" "replica" {
  depends_on           = [google_sql_database_instance.master]
  name                 = "mts-issue-test-replica-001"
  project              = data.google_project.project.project_id
  master_instance_name = "mts-issue-test-master-001"
  region               = "asia-northeast1"
  database_version     = "MYSQL_5_7"
  replica_configuration {
    failover_target = false
  }
  settings {
    tier              = "db-f1-micro"
    availability_type = "ZONAL"
    pricing_plan      = "PER_USE"
    disk_autoresize   = true
    disk_size         = 10

    backup_configuration {
      binary_log_enabled = true
    }

    database_flags {
      name  = "slave_parallel_workers"
      value = "3"
    }

    database_flags {
      name  = "slave_parallel_type"
      value = "LOGICAL_CLOCK"
    }

    database_flags {
      name  = "slave_pending_jobs_size_max"
      value = "536870912" # 512MB
    }
  }
}

Debug Output

when create new replica with binary logging (success)

https://gist.github.com/Climber22/09449433d0433104dfa7f8b74b04df42

when update replica (failed)

https://gist.github.com/Climber22/bc2abc80e52bb1bcb9e18f261180d6f7

Panic Output

Expected Behavior

When you run terraform apply, updating Cloud SQL completes without error.

A read replica instance with binary logging enabled is created.

Actual Behavior

When you run terraform apply again, updating Cloud SQL was failed with error message below.

│ Error: Error, failed to update instance settings for : googleapi: Error 400: Invalid request: Binary log must be disabled when backup is disabled or the instance must be a replica instance with a MySQL 5.7 or above version., invalid
│ 
│   with google_sql_database_instance.replica,
│   on cloudsql.tf line 35, in resource "google_sql_database_instance" "replica":
│   35: resource "google_sql_database_instance" "replica" {

Terraform tries to update read replica (mysql 5.7), so error message above looks wrong.
I wrote information about the reason for this confusing error message (Reason of this issue section)

Steps to Reproduce

When you firstly run terraform apply, Cloud SQL instance will be successfully created without error.
(Thanks for This PR, when create new one, PATCH method is explicitly called to enable binary_logging for replica instance )

Then, if you add some diff, and run terraform apply again, updating will be failed.

1. terraform apply

Important Factoids

Reason of this issue

I ask this issue for GCP support, and they said that “This is because the difference between PATCH and PUT method”.

In detail, PATCH method will evaluate diff between current status and request body, and apply only diff.
On the other hand, PUT method will try to overwrite all parameter with request body.

ref: https://cloud.google.com/sql/docs/mysql/admin-api/rest/v1beta4/instances/patch?hl=ja

Updates settings of a Cloud SQL instance. This method supports patch semantics.

In this case, when terraform try to update Cloud SQL by PUT method, there is no information to show “this is replica instance”.
To add this information, terraform should add instanceType params and specify it as “READ_REPLICA_INSTANCE“
ref: https://cloud.google.com/sql/docs/mysql/admin-api/rest/v1beta4/instances?hl=ja#SqlInstanceType

How to resolve this issue

In my opinion, adding instanceType field to google_sql_database_instance resource will solve this issue.
Additionally, If google_sql_database_instance has instanceType field, creation of read replica with binary logging enabled can be done without PATCH method ( I’ve not tried it, just my hypothesis )

References

• #0000