Installation manual

20
Value from 0 to 100. Influences overall spam score. Standard value is 100, i.e. in case of a positive
check the message is assigned the spam score of 100 and is evaluated as spam. Negative values
lower the overall spam score of a message.
DNSBL checks can have negative influence on server performance due to the fact that every domain/IP
address from the message body is checked against all defined DNSBL servers and every single check
requires processing a DNS server request. You can reduce the impact on system resources by deploying
a DNS cache server for this purpose. For the same reason the non-routable IP addresses (10.x.x.x,
127.x.x.x, 192.168.x.x) are also omitted from DNSBL checks.
Example 1:
dnsbl_list=ent.adbl.org
DNSBL check is performed against the ent.adbl.org server. If there is a positive, the message will be
assigned the default offset 100 (it will be marked as spam).
Example 2:
dnsbl_list=ent.adbl.org::60
DNSBL check is performed using the ent.adbl.org server. If the check is positive, the message will be
assigned an offset of 60 which increases its overall spam score.
Example 3:
dnsbl_list=bx9.dbl.com::85, list.dnb.org:127.0.0.4:35, req.gsender.org::-75
DNSBL check is performed using the defined servers (from left to right). If there is a positive check on
bx9.dbl.com, the offset of 85 will be added. If the check on list.dnb.org will be positive, giving a
response of 127.0.0.4 an offset of 35 will be used. No offset will be applied in cases of answers other
than 127.0.0.4. If a check is positive on req.gsender.org the spam score will be decreased by 75
points (negative value).
home_country_list
List of countries, that will be considered "home". Messages routed through a country not on this list
will be evaluated using more strict rules (higher spam score will be applied). Entry format for
countries is their two character code in compliance with ISO 3166.
home_language_list
List of preferred languages – i.e. languages that are the most used in your email messages. Such
messages will be evaluated using less strict rules (lower spam score). Entry format for languages is
their two character code in compliance with ISO 639.
custom_rules_list
Allows you to define custom lists of rules and store each list to an individual file. Each rule is stored
on a separate line in the file in the following format:
Phrase, Type, Confidence, CaseSensitivity
Phrase – Any text, must not contain commas (,).
Type – Can have the following values: SPAM, PHISH, BOUNCE, ADULT, FRAUD. If you enter other value
that those listed above, the SPAM value will be used automatically. SPAM defines phrases that occur in
classical spam messages (offers of goods and services). PHISH are phrases occurring in fraudulent
messages (phishing), that are aimed at extraction of confidential data (names, passwords, credit card
numbers, etc.) from users. BOUNCE are phrases used in automatic server responses - Non-Delivery
Notification (used when spoofing sender's address). ADULT represents phrases typical for messages
offering pornographic content. FRAUD stands for phrases used in fraudulent emails (scam) offering
suspicious banking operations (money transfers via your account etc.). A typical example of this spam
type is the so-called Nigerian spam.
Confidence – Value from 0 to 100. Defines the probability of the phrase to be member of a specific spam
category (listed above). If the Type PHISH has the Confidence 90, there is a very high probability of the
phrase being used in phishing messages. The higher the Confidence score, the bigger impact it exerts
on the overall spam score of the message. The Confidence value of 100 presents a special case, where
the message spam score will also be 100, i.e. message will be marked as 100% spam. Analogically, if
the value is 0, the message will be marked as not-spam.
CaseSensitivity – values 0 or 1. 0 meaning the phrase is case insensitive. 1 meaning the phrase is case
sensitive.
Examples:
replica, SPAM, 100, 0
Dear eBay member, PHISH, 90, 1
return to sender, BOUNCE, 80, 0