Pull Request to fix resource leaking in BroControl Python API
connDelete() is never called in broctl, which is generally not an issue if it is being used via the command line. However, given the Python interface for calling into broctl, if a service is written around broctl and any parallel events are sent (peerstatus, netstats, etc.) then a TCP connection will be maintained given a connDelete() is not called. After a fairly certain interval, resources become an issue and things fail.
Each time a broctl command is run that uses broccoli, a new TCP connection
is established. The connection is not closed until one either exits from broctl,
or stops the Bro nodes.
Ideally, it would reuse the TCP connection, but at least your fix is better than the current