[Bug]: aws_emr_cluster Inappropriate value for attribute "configurations": string required.

[devlatte]

Terraform Core Version

1.9.8

AWS Provider Version

5.69.0

Affected Resource(s)

While working with EMR, I followed the Terraform guide but encountered an error:
Incorrect attribute value type about configurations in aws_emr_cluster.

Could you please confirm this?
I would prefer to use this option as it's more readable than configurations_json.

Expected Behavior

working well

Actual Behavior

Incorrect attribute value type occured

Relevant Error/Panic Output Snippet

╷
│ Error: Incorrect attribute value type
│
│   on adhoc_spark.tf line 47, in resource "aws_emr_cluster" "emr_adhoc_spark":
│   47:   configurations = [
│   48:     {
│   49:       classification = "spark-defaults"
│   50:       properties = {
│   51:         "spark.dynamicAllocation.enabled"   = "false"
│   52:         "spark.executor.memory"             = "4g"
│   53:         "spark.emr.default.executor.memory" = "4g"
│   54:         "spark.sql.session.timeZone"        = "Asia/Seoul"
│   55:       }
│   56:     },
│   57:     {
│   58:       classification = "spark-hive-site"
│   59:       properties = {
│   60:         "hive.metastore.client.factory.class" = "com.amazonaws.glue.catalog.metastore.AWSGlueDataCatalogHiveClientFactory"
│   61:       }
│   62:     }
│   63:   ]
│
│ Inappropriate value for attribute "configurations": string required.
╵

Terraform Configuration Files

resource "aws_emr_cluster" "emr_adhoc_spark" {
  ...
  configurations = [
    {
      classification = "spark-defaults"
      properties = {
        "spark.dynamicAllocation.enabled"   = "false"
        "spark.executor.memory"             = "4g"
        "spark.emr.default.executor.memory" = "4g"
        "spark.sql.session.timeZone"        = "Asia/Seoul"
      }
    },
    {
      classification = "spark-hive-site"
      properties = {
        "hive.metastore.client.factory.class" = "com.amazonaws.glue.catalog.metastore.AWSGlueDataCatalogHiveClientFactory"
      }
    }
  ]
}

Steps to Reproduce

1. write emr configuration
2. execute terraform plan

Debug Output

No response

Panic Output

No response

Important Factoids

No response

References

https://github.com/hashicorp/terraform-provider-aws/blob/main/internal/service/emr/cluster.go#L308

Would you like to implement a fix?

None

[github-actions]

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]

Hey @devlatte 👋 Thank you for taking the time to raise this! I took a look over the documentation and found it a bit confusing as well, so I took a look through the relevant parts of the source code and the test cases and found that the expectation for the configurations argument is that the string will either point to a local file or to a URL containing the configuration data. It definitely seems like the documentation for that argument could use a bit of touching up to make that more clear.

On the other hand, your sample configuration could be made to work with configurations_json by using jsonencode():

resource "aws_emr_cluster" "emr_adhoc_spark" {
  # ...other config...

  configurations_json = jsonencode([
    {
      classification = "spark-defaults"
      properties = {
        "spark.dynamicAllocation.enabled"   = "false"
        "spark.executor.memory"             = "4g"
        "spark.emr.default.executor.memory" = "4g"
        "spark.sql.session.timeZone"        = "Asia/Seoul"
      }
    },
    {
      classification = "spark-hive-site"
      properties = {
        "hive.metastore.client.factory.class" = "com.amazonaws.glue.catalog.metastore.AWSGlueDataCatalogHiveClientFactory"
      }
    }
  ])
}

[devlatte]

Hi @justinretzolk
Managing it with a file is a good option, and the jsonencode() method seems useful too.
Both methods seem like better options than a JSON string :)
I'll try applying it.
Thanks for the reply!