### Summary Running the script `delete_old_cluster.sh` doesn't remove the *old* cluster, but the running(!) cluster. ### Steps to reproduce Run the script. ### What is the current *bug* behavior? The fact that the data is removed without warning and without checks, from a path that holds production data in some configurations (all?). ### What is the expected *correct* behavior? There should at least have been safety checks in the script, e.g. by moving the cluster aside, restarting postgresql, asking the user if the site is still live, etc.