Inconsistent job placement response with constraints on system jobs

Nomad version

1.6.1

Operating system and Environment details

AWS EC2, Ubuntu 22.04

Issue

When deploying a system type job with a specific constraint filter, the job appears to be placed correctly on all targeted nodes. However, the CLI and UI indicate a placement failure. This behavior is inconsistent and misleading, as the job is actually being applied as expected.

This issue leads to confusion and potential misinterpretation of the job's deployment status. It also affects automation workflows that rely on CLI exit codes.

Reproduction steps

Define a system type job with a constraint filter.
Deploy the job.
Observe that, despite successful placement on all filtered nodes, the CLI exits with a non-zero code and the UI shows a placement failure.

Expected Result

The expectation for a system type job with a constraint filter is to have it placed on all nodes that meet the constraint criteria without any errors or misleading failure messages.

Actual Result

The job gets placed on the appropriate nodes, but the CLI and UI report a placement failure, which is not accurate.

Job file (if appropriate)

Simplified example:

job "my-app" {
  type = "system"

  group "my-app" {
    count = 1

    task "my-app" {
      driver = "exec"

      artifact {
        source = "s3::https://s3-ap-south-1.amazonaws.com/app/path/binary.tar.gz"
      }

      # Run the program.
      config {
        command = "$${NOMAD_TASK_DIR}/app.bin"
      }

      constraint {
        attribute = "${meta.ec2_nomad_client}"
        operator  = "="
        value     = "my-app"
      }
    }
  }
}

The meta.ec2_nomad_client exists in all client nodes, but they have different values. We use this tag to determine which job gets placed in which node.

hashicorp / nomad