rutujachaudhari124 commented 1 year ago

Terraform Core Version

0.12.31

AWS Provider Version

3.70.0

Affected Resource(s)

aws_rds_global_cluster

Expected Behavior

Subsequent apply should show no changes and no force replacement.

Actual Behavior

The subsequent apply shows a diff for the database_name parameter since this value is inherited from source cluster configuration and the applied configuration contains N/A value for this parameter.

-/+ resource "aws_rds_global_cluster" "this" {
      ~ arn                          = "************" -> (known after apply)
      - database_name                = "database04" -> null # forces replacement
        deletion_protection          = false
      ~ engine                       = "aurora" -> (known after apply)
      ~ engine_version               = "5.6.mysql_aurora.1.22.5" -> (known after apply)
        force_destroy                = true
        global_cluster_identifier    = "***************"
      ~ global_cluster_members       = [
          - {
              - db_cluster_arn = "******************"
              - is_writer      = true
            },
        ] -> (known after apply)
      ~ global_cluster_resource_id   = "**********" -> (known after apply)
      ~ id                           = "*************" -> (known after apply)
        source_db_cluster_identifier = "****************"
      ~ storage_encrypted            = true -> (known after apply)
    }

Plan: 1 to add, 0 to change, 1 to destroy.

Relevant Error/Panic Output Snippet

Does not occur any error while terraform apply, but force replacement is occurred for database_name parameter
database_name                = "database04" -> null # forces replacement

Terraform Configuration Files


resource "aws_rds_cluster" "primary" {
    apply_immediately                   = true
    availability_zones                  = ["ap-south-1a", "ap-south-1b", "ap-south-1c"]
    backtrack_window                    = 0
    backup_retention_period             = 1
    cluster_identifier                  = "****************"
    copy_tags_to_snapshot               = true
    database_name                       = "database04"
    deletion_protection                 = false
    enable_http_endpoint                = false
    enabled_cloudwatch_logs_exports     = ["error", "slowquery"]
    engine_mode                         = "provisioned"
    engine_version                      = "5.6.mysql_aurora.1.22.5"
    engine                              = "aurora"
    final_snapshot_identifier           = "******************"
    iam_database_authentication_enabled = false
    master_password                     = "*******"
    master_username                     = "dbadmin"
    port                                = 5432
    preferred_backup_window             = "03:30-05:00"
    preferred_maintenance_window        = "sun:19:00-mon:00:00"
    skip_final_snapshot                 = true
    storage_encrypted                   = false

}

data "aws_rds_cluster" "clusterName" {
  cluster_identifier = "****************"

}

resource "aws_rds_global_cluster" "example" {
  source_db_cluster_identifier = data.aws_rds_cluster.clusterName.arn
    global_cluster_identifier = "******************"
    deletion_protection       = false
    force_destroy             = true
}

Steps to Reproduce

terraform apply terraform plan

Debug Output

No response

Panic Output

No response

Important Factoids

No response

References

This github issue is similar issue for 'storage_encrypted' parameter in this same resource. https://github.com/hashicorp/terraform-provider-aws/issues/15177

Would you like to implement a fix?

None

github-actions[bot] commented 1 year ago

Community Note

Voting for Prioritization

Please vote on this issue by adding a 👍 reaction to the original post to help the community and maintainers prioritize this request.
Please see our prioritization guide for information on how we prioritize.
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.

Volunteering to Work on This Issue

If you are interested in working on this issue, please leave a comment.
If this would be your first contribution, please review the contribution guide.

justinretzolk commented 1 year ago

Hey @rutujachaudhari124 👋 Thank you for taking the time to raise this! So that we have the information needed to look into this, can you supply a sample Terraform configuration as well?

Kristin0 commented 1 year ago

have you tried to specify database_name it implicitly ?

aws_rds_global_cluster = {
  ***************** = {
    global_cluster_identifier    = "*************"
    deletion_protection          = false
    database_name                = "database04"
    force_destroy                = true
    source_db_cluster_identifier = "*************"
  }
}

rutujachaudhari124 commented 1 year ago

have you tried to specify database_name it implicitly ?

aws_rds_global_cluster = {
  ***************** = {
    global_cluster_identifier    = "*************"
    deletion_protection          = false
    database_name                = "database04"
    force_destroy                = true
    source_db_cluster_identifier = "*************"
  }
}

@Kristin0 We have tested this case where we have added the database_name as "database04", in this case we got an error that the database_name should not be specified as it will be inherited from source cluster. The error is mentioned below:

rutujachaudhari124 commented 1 year ago

@justinretzolk Updated the terraform configuration.

Kristin0 commented 1 year ago

@justinretzolk any updates?

vitali-miadzvedzki commented 1 year ago

Looking into similar issue mentioned here #15177 it seems the database_name parameter requires to have computed=true (regional has it: https://github.com/hashicorp/terraform-provider-aws/blob/v4.46.0/internal/service/rds/cluster.go#L122, global does not: https://github.com/hashicorp/terraform-provider-aws/blob/v4.46.0/internal/service/rds/global_cluster.go#L48)

Kristin0 commented 1 year ago

same issue with these versions:

Terraform Core Version
v1.3.3

AWS Provider Version
v4.46.0

ollieanwyll commented 1 year ago

Having this same issue. Has anyone had any luck?

CGandhi0 commented 1 year ago

We are have a similar issue with aurora global cluster for postgres built out of a snapshot. With is_promoted_global_cluster = true, global cluster gets correctly setup with database_name that is identical to regional cluster but upon subsequent apply instead of detecting no changes to infrastructure - terraform is noticing database_name getting updated from current database_name to null which forces replacement of global cluster.

-/+ resource "aws_rds_global_cluster" "global_cluster_promoted" {
      ~ arn                          = "arn:aws:rds::******:global-cluster:****" -> (known after apply)
      - database_name                = "*****" -> null # forces replacement
      ~ engine                       = "aurora-postgresql" -> (known after apply)
      ~ engine_version               = "13.6" -> (known after apply)
      ~ engine_version_actual        = "13.6" -> (known after apply)
      ~ global_cluster_members       = [
          - {
              - db_cluster_arn = "arn:aws:rds:us-east-1:*******:cluster:********"
              - is_writer      = false
            },
          - {
              - db_cluster_arn = "arn:aws:rds:us-west-2:*******:cluster:********"
              - is_writer      = true
            },
        ] -> (known after apply)
      ~ global_cluster_resource_id   = "cluster-3d909f38d460f47e" -> (known after apply)
      ~ id                           = "********" -> (known after apply)
      ~ storage_encrypted            = true -> (known after apply)
        # (4 unchanged attributes hidden)
    }

Plan: 1 to add, 0 to change, 1 to destroy.

If you try to apply it fails with following error

Error: error removing RDS DB Cluster (arn:aws:rds:us-west-2:******:cluster:********) from Global Cluster (******): InvalidParameterValue: Can not remove writer cluster when there are other clusters
│   status code: 400, request id: 1c891562-958e-4ae0-a7ce-3a6cc0dcdb22

If you don't have a secondary cluster or destroy secondary cluster and run apply it will still recreate global cluster but at this time apply doesn't fail and it will succesfully complete. However, re-running apply still show same issue about global cluster needs to be recreated.

CGandhi0 commented 1 year ago

Any updates...this is blocking all subsequent updates to the project that includes global cluster spread across multiple regions.

CGandhi0 commented 1 year ago

@justinretzolk - any idea on when this will get address? It is blocking all updates to the project that include other resources.

cstclair-tunein commented 1 year ago

+1 to this. We are currently experiencing this bug.

justinretzolk commented 1 year ago

Hey y'all 👋 Thank you for checking in on this! Unfortunately I can't provide an ETA on when this will be looked into due to the potential of shifting priorities. We prioritize by count of :+1: reactions and a few other things (more information on our prioritization guide if you're interested).

mnebot commented 1 year ago

Hi all,

I faced the same issue, I used this workarround, hope that could work also with you

resource "aws_rds_global_cluster" "example" {
  ...
  lifecycle {
    ignore_changes = [database_name]
  }
}

CGandhi0 commented 1 year ago

I used ignore_changes to database_name as workaround to mitigate issue. Forgot to mention it here.

hashicorp / terraform-provider-aws

[Bug]: Subsequent apply forces global cluster recreation when source cluster's database_name is specified #28187

Terraform Core Version

AWS Provider Version

Affected Resource(s)

Expected Behavior

Actual Behavior

Relevant Error/Panic Output Snippet

Terraform Configuration Files

Steps to Reproduce

Debug Output

Panic Output

Important Factoids

References

Would you like to implement a fix?

Community Note