RDS shared_preload_libraries update inconsistent.

[AndrewNeudegg]

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 other comments that do not add relevant new information or questions, 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

Terraform Version

Terraform v0.11.14

Affected Resource(s)

RDS Parameter groups, shared_preload_libraries.

• aws_XXXXX

Terraform Configuration Files

Using: https://github.com/terraform-aws-modules/terraform-aws-rds

module "db_1" {
  source = "terraform-aws-modules/rds/aws"
  version = "~> 1.0"
  
  identifier = "db_1"

  engine            = "postgres"
  engine_version    = "11.5"
  instance_class    = "${var.er_instance_type}"
  allocated_storage = "${var.er_storage}"
  storage_encrypted = true
  storage_type      = "gp2"
  performance_insights_enabled = true
  apply_immediately = true

  allow_major_version_upgrade = true

  name = "db_1"

  # NOTE: Do NOT use 'user' as the value for 'username' as it throws:
  # "Error creating DB Instance: InvalidParameterValue: MasterUsername
  # user cannot be used as it is a reserved word used by the engine"
  username = "er_admin"

  password = "${var.db_1_password}"
  port     = "5432"

  vpc_security_group_ids = ["${aws_security_group.databases.id}"]
  multi_az               = true

  maintenance_window = "Tue:00:00-Tue:03:00"
  backup_window      = "03:00-06:00"

  # disable backups to create DB faster
  backup_retention_period = 7

  tags = {
    Owner = "chris.brook"
  }

  enabled_cloudwatch_logs_exports = ["postgresql", "upgrade"]

  # DB subnet group
  subnet_ids = ["${data.aws_subnet_ids.private.ids}"]

  # DB parameter group
  create_db_parameter_group = true

  # parameter_group_name = "db_1"
  parameter_group_description = "db_1 specific parameters"
  family                      = "postgres11"

  # note that the apply method must be determined from the parameter type. static parameters can only be
  # applied after reboot, whereas dynamic parameters can be applied immediately. the implicit default is
  # to apply immediately, so any static parameters must have an apply method of "pending-reboot". it's 
  # unclear why terraform can't infer this information when the apply method isn't specified - probably
  # something web-scale related..
  parameters = [
    {
      name = "random_page_cost"

      value = 1.1
    },    
    {
      name = "rds.force_ssl"

      value = 1
    },
    {
      name = "auto_explain.log_min_duration"

      value = 5000
    },
    {
      name = "auto_explain.log_nested_statements"

      value = 1
    },
    {
      name = "log_min_duration_statement"

      value = 5000
    },
    {
      name = "log_statement"

      value = "none"
    },
    {
      name = "pg_hint_plan.debug_print"

      value = "on"
    },
    {
      name = "pg_hint_plan.enable_hint"

      value = "on"
    },
    {
      name = "pg_stat_statements.max"

      value = 10000,

      apply_method = "pending-reboot"
    },
    {
      name = "pg_stat_statements.track"

      value = "all"
    },
    {
      name = "shared_preload_libraries"

      value = "${join(",", "${var.shared_preload_libraries}")}"

      apply_method = "pending-reboot"
    },
  ]

  use_parameter_group_name_prefix = true
  options                         = []

  major_engine_version = "11.5"

  # Snapshot name upon DB deletion
  final_snapshot_identifier = "db_1"

  # Database Deletion Protection
  deletion_protection = true

  copy_tags_to_snapshot = true

  create_monitoring_role              = false
  iam_database_authentication_enabled = true
  monitoring_interval                 = 5
  monitoring_role_arn                 = "${aws_iam_role.enhanced_monitoring.arn}"
}

With the configuration:

variable "shared_preload_libraries" {
  type    = "list"
  default = [ "pg_stat_statements", "auto_explain", "pg_hint_plan" ]
}

Debug Output

Panic Output

Expected Behavior

When values are added and removed from the configuration array, the changes should be properly reflected in shared_preload_libraries.

Actual Behavior

In the instance that the values in configuration diverge from the values in shared_preload_libraries the plan looks like it executes properly but shared_preload_libraries contains only the pg_stats_statements:

In the instance that the value pg_stats_statements is not in the configuration, the values in shared_preload_librariesupdate properly.

In the instance that the value of shared_preload_libraries is the same as the value in configuration -- no change occurs.

Steps to Reproduce

1. Change the configuration from:

variable "shared_preload_libraries" {
  type    = "list"
  default = [ "pg_stat_statements", "pg_hint_plan" ]
}

to

variable "shared_preload_libraries" {
  type    = "list"
  default = [ "pg_stat_statements", "auto_explain", "pg_hint_plan" ]
}

And check the value of shared_preload_libraries.

Important Factoids

References

• aws_rds_cluster_parameter_group change in a parameter reset it #11846 -- potentially related?

[caleb15]

This is also a issue with terraform version 0.12.29.

[caleb15]

Our configuration is like so:

  parameter {
    name         = "shared_preload_libraries"
    value        = "pg_stat_statements,auto_explain"
    apply_method = "pending-reboot"
  }

[caleb15]

Oh no, it looks like some other params might be affected too 😬

auto_explain.log_buffers, auto_explain.log_triggers, auto_explain.log_verbose, auto_explain.log_nested_statements, auto_explain.log_format, and auto_explain.log_min_duration all have the same problem in some of our VPC's.

[justinretzolk]

Hi all 👋 Thank you for taking the time to file this issue, and for the discussion so far. Given that there's been a number of Terraform and AWS provider releases since the last update, can anyone confirm whether they're still experiencing this issue?

[phil-hildebrand]

I'm seeing this issue in terraform version 0.15.3

[justinretzolk]

Hey @phil-hildebrand 👋 What version of the AWS Provider are you using where you're experiencing this?

[phil-hildebrand]

The aws provider version was 2.70.0.

[justinretzolk]

Hey @phil-hildebrand 👋 Version 2.70.0 was released quite a while ago, and there've been many releases since. Are you able to test with a newer version of the provider to determine whether you experience the same issues?

[phil-hildebrand]

We are in the process of upgrading to 3.75.1, I'll test that this week and note if it addresses it.

[phil-hildebrand]

We are in the process of upgrading to 3.75.1, I'll test that this week and note if it addresses it.

We verified that upgrading fixed the issue

[justinretzolk]

Thanks for the update! Given that it seems this was indeed fixed in later versions, I'll get this issue closed out. If anyone experiences this in the future, please open a fresh issue so we can take another look.

[github-actions]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.