Requirement: Given the requirement for compute resources capable of handling tasks lasting over 15 minutes, we determined that AWS Batch was the optimal solution.

Problem: We support versioning of our application. However, AWS Batch doesn’t support versioning like AWS Lambdas. Though, it creates a new JobDefinition with a new incremental revision after each deployment.

Example:

Here, the revision = 58.

However, it marks the previous revision of the JobDefinition as inactive and eligible for deletion automatically.

Like:

When we try to submit Job on any of these previous revision using aws sdk.

Like:

SubmitJobRequest submitJobRequest =
    SubmitJobRequest.builder()
        .jobName("sample")
        .jobQueue("sample")
        .jobDefinition("sample:"+ revision)
        .containerOverrides(
            ContainerOverrides.builder()
                .command(
                    "java",
                    "-cp",
                    "app.jar",
                    "-Dspring.main.web-application-type=none",
                    "-Dloader.main=com.BatchJob",
                    "org.springframework.boot.loader.PropertiesLauncher")
                .build())
        .build();

SubmitJobResponse submitJobResponse = batchClient.submitJob(submitJobRequest);

Exception:

JobDefinition arn:aws:batch:eu-west-1:123456789123:job-definition/sample:57 is not in ACTIVE status. (Service: Batch, Status Code: 400, Request ID: bd6199a1-3978-458c-a84e-19de4c1acb62)

Answers already read:

I’ve read below answer on the Stack Overflow but none of the answers answered my question:

1. aws Batch Jobdefinition Cloudfromation error

2. aws lambda cannot start step functions in eu-west-1

3. AWS batch job submit error-Batch.ClientException

Any help would be very appreciated!

MORE DETAIL:

We are using terraform as Infrastructure as Code (IaC) and my aws_batch_job_definition look like:

resource "aws_batch_job_definition" "test" {
  name = "sample"
  type = "container"

  platform_capabilities = [
    "FARGATE",
  ]

  container_properties = jsonencode({
    command    = ["echo", "test"]
    image      = "${var.ecr_url_primary}:${local.version}-sample"
    jobRoleArn = aws_iam_role.job_task_role.arn

    fargatePlatformConfiguration = {
      platformVersion = "LATEST"
    }

    resourceRequirements = [
      {
        type  = "VCPU"
        value = "1"
      },
      {
        type  = "MEMORY"
        value = "3072"
      }
    ]
    executionRoleArn = aws_iam_role.execution_role.arn
    environment = [
      {
        name = "SPRING_PROFILES_ACTIVE"
        value = "${tostring(var.env)}"
      }
    ]
  })
  provider = "aws"
}