Shared VPS Dedicated WP Professional Plus
This article describes our advanced junk email filtering options. If you haven't already, please first read the Junk Email Filtering Overview.
If you've already enabled junk email filtering and just need to set up your email reader, please read Junk Email Filtering And Your Email Reader.
DNSBLs
A DNSBL, or DNS Blackhole List, is a service that maintains a list of sites believed to be responsible for large volumes of junk email. You can choose to have your email checked against DNSBLs for determining a SpamAssassin junk email score. SpamAssassin uses a scoring system to determine which email messages are junk and which are legitimate.
You can enable the DNSBLs that SpamAssassin uses for junk email scoring by following the steps below:
- Log in to the Account Control Center (https://my.pair.com).
- Click E-Mail.
- Click Manage Junk E-Mail Filter Settings.
- Click Junk E-Mail Filtering Options for Advanced Users.
- Select Use SpamAssassin DNSBLs for junk e-mail filtering scoring.
- Click Commit Changes.
Please do not enable this option unless you are completely familiar with DNSBLs and the implications of using this option. This option is not recommended for inexperienced users.
Invaluement DNSBL
Pair Networks also grants you access to a DNSBL called Invaluement DNSBL (visit http://www.invaluement.com/ for more information) to reject junk email messages and to block those addresses that send junk email.
You can enable Invaluement DNSBL by following the steps below:
- Log in to the Account Control Center (https://my.pair.com).
- Click E-Mail.
- Click Manage Junk E-Mail Filter Settings.
- Click Junk E-Mail Filter Settings for Advanced Users.
- Select Use Invaluement DNSBL to reject e-mail that Invaluement flags as junk.
- Click Commit Changes.
Please do not enable Invaluement DNSBL unless you are completely familiar with DNSBLs.
Bayesian Filtering
A Bayesian Filter works by studying the words found in normal and junk email. It uses this analysis to determine the probability that future emails are junk by looking at the words they contain.
Bayesian Filtering can also be enabled in the Email Management section of the Account Control Center on the "Junk E-Mail Filtering Settings for Advanced Users" page.
The filter will not begin to affect the scores of scanned emails until it has already learned from 200 junk emails and 200 regular emails. Emails with especially high or low scores are learned automatically, but the process may take several weeks before the filter becomes active.
If you have files containing junk or regular emails you wish the filter to learn, you may connect via SSH and use the sa-learn command:
To learn a file of junk email messages, use:
sa-learn --spam --mbox filename
To learn a file of regular email messages, use:
sa-learn --ham --mbox filename
For a complete list of options, run:
sa-learn --help
Email files being learned must be less than 30MB in size. Be sure to remove any email that has already been learned before sending the file to sa-learn -- these email messages will not be learned again but will still count toward the Bayesian Filtering limit.
The system transfers email files to the email filtering server for Bayesian Filtering learning every 20 minutes.
Custom Preferences File
In addition to the SpamAssassin preferences file created by your choices in the Account Control Center, you can keep a preference file in your home directory. This file should be named user_prefs and located in the directory /usr/home/username/.spamassassin/. The maximum size for your customer preferences file is 15k.
Useful uses for this file include:
- Setting a custom score for mail to be considered junk:
required_score 9.5
- Disabling certain rules:
score SOME_RULE 0
- Whitelisting/blacklisting addresses:
whitelist_from [email protected] blacklist_from [email protected]
This is especially useful for uploading large lists or dynamically maintaining the list.
NOTE: When working with the raw configuration file, * must be used as a wild-card character. For instance, use *@example.com to block all mail from example.com.