As described on the files page this
file is identical in function to bad_hosts
except that it is intended to be downloaded from a central resource rather than
maintained locally. If you choose not to use it, simply ignore these
instructions, your Cleanfeed will work fine without it.
How It Works
The bad_hosts_central file is generated automatically by a process that runs on
news.mixmin.net. It scans the logfiles produced by its local Cleanfeed and
identifies problem hosts that are generating lots of rejected messages. If a
threshold number of rejected messages is crossed in a day, the host gets added
to the bad_hosts_central file for a period of time (currently 3 days). If
during that period the host continues to exceed the reject threshold, the period of time
it stays listed keeps rolling forward.
Downloading the file
The bad_hosts_central file is available via FTP or HTTP from
www.mixmin.net/cleanfeed/bad_hosts_central
The recommended method for obtaining it is via a cron entry like the following:-
Make sure that bad_rate_reload is set at a
sensible figure for the server you operate or the downloaded file may not be read by
Cleanfeed for a long period of time. Conversely don't set it too low or Cleanfeed
will be forever rereading its bad_* files which isn't good for performance.
NNTP-Posting-Host Issue
Some News providers choose to put information other than the posting host in
this header. In some instances it's an encrypted or hashed version of the
actual posting host which is completely compatible with the process that
generates the bad_hosts_central file. In one instance that I'm aware of, a
major provider is putting their organisation name in the NNTP-Posting-Host
header. This will result in them being frequently being listed as a bad host
as every message originating from their service appears to come from a single
source. Please don't misuse the headers in this manner!