There are a number of variables set in the various lib/ files. Most of the ones you would want to change are settable from the CreateBroker script. One useful file to know about is lib/default-ignore. This file lists regular expressions for names of files to avoid replicating. The lib/default-ignore file included in the Harvest distribution contains some regular expressions used to avoid files such as temporary copies of editor sessions, core dumps, and log files.
Another useful customization is to create an access control list. To do this, edit group.conf, to point to a file listing class A/B/C network numbers to allow or deny. An example follows:
;; ;; A paranoid group that only talks to known hosts. ;; (:group-define (:group-name"group1") (:allow-site "128.125.0.0") ; allow access from USC (:allow-site "128.138.0.0")) ; allow access from UColo - Boulder ;; ;; A permissive group that disallows known trouble sites. ;; (:group-define (:group-name "group2") (:deny-site "18.0.0.0")) ; deny access to MIT hackers!