puppetlabs/puppet · error · Puppet::Error

launchctl list failed to return any data.

Error message

launchctl list failed to return any data.

What it means

The launchd provider's job_list class method runs launchctl list and immediately raises if the command produced nil output (execute returned nothing at all, distinct from an empty-but-successful string). A separate rescue converts Puppet::ExecutionFailure into 'Unable to determine status'. The nil check exists because the code then calls output.split, which would otherwise raise NoMethodError.

Source

Thrown at lib/puppet/provider/service/launchd.rb:190

        if by_label.has_key? label
          { label => by_label[label] }
        else
          raise Puppet::Error, "Unable to find launchd plist for job: #{label}"
        end
      end
    else
      # caller wants the whole map
      by_label
    end
  end

  # This status method lists out all currently running services.
  # This hash is returned at the end of the method.
  def self.job_list
    @job_list = Hash.new
    begin
      output = launchctl :list
      raise Puppet::Error, "launchctl list failed to return any data." if output.nil?

      output.split("\n").each do |line|
        @job_list[line.split(/\s/).last] = :running
      end
    rescue Puppet::ExecutionFailure => e
      raise Puppet::Error.new("Unable to determine status of #{resource[:name]}", e)
    end
    @job_list
  end

  # Read a plist, whether its format is XML or in Apple's "binary1"
  # format.
  def self.read_plist(path)
    Puppet::Util::Plist.read_plist_file(path)
  end

  # Read overrides plist, retrying if necessary
  def self.read_overrides

View on GitHub (pinned to e227c27540)

Solutions

1. Run launchctl list as the same user puppet uses and confirm it prints lines
2. Stop forcing provider => 'launchd' on non-macOS nodes; let the provider confine to darwin pick it
3. Check which launchctl resolves: which -a launchctl, and remove wrappers that eat stdout
4. Repair the OS binary / reimage broken hosts rather than suppressing the error

out=$(launchctl list 2>&1)
if [ -z "$out" ]; then echo "launchctl returned no data — repair OS or stop forcing launchd provider"; fi

When it happens

Trigger: Calling self.job_list (status/prefetch paths of the launchd provider) in an environment where the launchctl binary resolves but returns nothing — a broken/shadowed launchctl, an execution wrapper swallowing output, or a non-macOS sandbox where the provider was forced via provider => 'launchd'.

Common situations: Running puppet inside CI containers or over a non-macOS test host with the provider forced for spec runs; macOS images where launchctl is damaged or SIP/permissions block it; wrapper scripts in PATH intercepting system binaries.

Related errors

• Unable to find launchd plist for job: #{label}
• Unable to read overrides plist, too many attempts
• Unable to parse launchd plist at path: #{job_path}
• puppet.plans/invalid-name
• Invalid #{field} given for user #{user_name}

AI-assisted analysis of puppetlabs/puppet@e227c27540 (2026-08-21). Data as JSON: /api/errors/fca65c124ef53ca6. Report an issue: GitHub.