analyze_cruft.sh

#!/bin/bash
# This script is designed to give an overview of which files are present
# in a directory which Puppet does not manage. It isn't perfect, but it
# works reasonably well.
# Specifically, it doesn't understand whether a directory contains
# subdirectories that shouldn't be purged, and it doesn't query the Puppet
# fileserver.

set -o errexit
set -o nounset
set -o pipefail

hostname=$(hostname -f)
search=${1%/}

existing=$(mktemp -t "analyze_cruft.XXXXXX")
find "${search}" -maxdepth 1 | sort > "${existing}"
catalog=$(mktemp -t "analyze_cruft.XXXXXX")
jq -r '.data.resources[] | select(.type == "File" and .ensure != "absent") | .parameters.path // .title' "/var/lib/puppet/client_data/catalog/${hostname}.json" \
  | grep '^'"$search" | sort > "${catalog}"
join -v1 "${existing}" "${catalog}"
rm "${existing}" "${catalog}"