Right now, puppet adds a config file to the directory in ores deployments in target which contain very basic setup such as redis but once we deploy via scap3, it creates another directory (in /srv/deployment/ores/deploy-cache/#hash) and makes the ores down until puppet makes the file again. After lots of talk we agreed to move the config file(s) to /etc/ores/ and make ores read two directories instead of one.