Make SPAM ASSASSIN work for you... - Page 9

**Bigjohn** · September 7th, 2006, 2:36 PM

yes, you can and should manually adjust scores... forged_rcvd_helo is one that I bumped up from the default of 1.5 or so to 3, for example.

kavaXtreme · September 10th, 2006, 4:05 PM

Bigjohn,

Thanks a ton for the tutorial! I've done a lot of reading about SpamAssassin, but it's hard to find much of use for Cpanel users. Your info is the the most helpful I've found anywhere on the web.

A couple of questions:
1. Could you add a section to the tutorial about which rules to modify and how to do it?
2. A significant amount of all spam I receive comes to addresses I don't use (uucp, majordomo, root, etc.). I could set up filters to discard those e-mails, but since they ONLY receive spam, they seem like ideal learning material for Bayes. Is there a way to set up Bayes to learn from them instead of using the previously mentioned method?

**Bigjohn** · September 10th, 2006, 10:41 PM

Kava -

thanks.

There is another more recent thread here in which Cowboy and I discussed and created an Exim filter. The filter can be configured to do a lot of things - for example, automatically move email to 'non existant' email addresses into your mailtrap account, where you can then train it as spam.

The key with rules and/or especially spam assassin is that they don't really look at the TO address - they should analyze the content.

So, if these emails are coming to MORE THAN ONE email address, you're learning against them already.

kavaXtreme · September 11th, 2006, 2:42 PM

For others reading this thread, you'll probably want to check out the following:

Updated script (auto-delete old spam):
Anyway to stop SPAM?

Exim filter "tutorial" (mentioned in Bigjohn's post above):
Whitelist based on subject line?

**Bigjohn** · September 11th, 2006, 3:10 PM

Yeah, you don't want to autodelete old spam...

If your bayes filter gets hosed, you want to have at least 500 spam messages available to re-train it quickly. Yes, I know you'll get 500 spam's in a couple of days, but they'll all end up in your inbox!

John

kavaXtreme · September 11th, 2006, 4:19 PM

Sorry. I should have clarified: it doesn't auto-delete all spam, just spam that has been in the mailbox for more than the specified number of days. So you will always have a healthy dose of fresh spam in the mailbox, but spam that's been sitting around for x number of days will be deleted.

kavaXtreme · September 14th, 2006, 6:54 PM

If you're like me and are not on a dedicated server, you might be interested in a reply I recently got from support: "We are sorry to say that Razor, Pyzor and DCC are not allowed." Also, you can't use custom rules (which is a major bummer for me, but I do understand there are security considerations).

Still, Bayes learning and RBL checking combined with mail filters has already made a HUGE difference for me. (And I'm not even using the script with Bayes... yet.)

grokdesigns · October 19th, 2006, 11:11 AM

Can anybody tell me why my mail isn't being processed? I get cron job reports as follows:

Code:

Learning SPAM
Processing /home/[name]/mail/grokdesigns.com/chris/.SPAM
Learned tokens from 0 message(s) (0 message(s) examined)
rm: cannot remove `/home/[name]/mail/grokdesigns.com/chris/.SPAM': Is a directory
 
Learning HAM
Processing /home/[name]/mail/grokdesigns.com/chris/.HAM
Learned tokens from 0 message(s) (0 message(s) examined)
rm: cannot remove `/home/[name]/mail/grokdesigns.com/chris/.HAM': Is a directory
Done

These are the correct directory names, but no messages are being processed...and I'm confused as to why. I have 132 messages in .SPAM and 69 in .HAM. I also don't understand why it's trying to remove the directory instead of messages.

**Bigjohn** · October 19th, 2006, 12:16 PM

try a trailing slash?

September 7th, 2006, 2:36 PM	#73 (permalink)
Bigjohn minor deity Super #1 Joined in Apr 2004 Lives in Georgia Hosted on XEON 7,340 posts Gave thanks: 23 Thanked 93 times	yes, you can and should manually adjust scores... forged_rcvd_helo is one that I bumped up from the default of 1.5 or so to 3, for example. __________________ Proud to be a Surmunity Mod! XEON PASS60 PASS61 Make a fundamental difference! My Sites: Curious about Brewing Beer? Join the community! >>>>> Some Change is GOOD! Keep your paycheck! Support the Fair Tax Get into an Art museum Victorian London It's your brain -*ON WEB* - mybrainhost.com (under development) What SHOULD Government do? Much Less than it Does!

September 10th, 2006, 10:41 PM	#75 (permalink)
Bigjohn minor deity Super #1 Joined in Apr 2004 Lives in Georgia Hosted on XEON 7,340 posts Gave thanks: 23 Thanked 93 times	Kava - thanks. There is another more recent thread here in which Cowboy and I discussed and created an Exim filter. The filter can be configured to do a lot of things - for example, automatically move email to 'non existant' email addresses into your mailtrap account, where you can then train it as spam. The key with rules and/or especially spam assassin is that they don't really look at the TO address - they should analyze the content. So, if these emails are coming to MORE THAN ONE email address, you're learning against them already. __________________ Proud to be a Surmunity Mod! XEON PASS60 PASS61 Make a fundamental difference! My Sites: Curious about Brewing Beer? Join the community! >>>>> Some Change is GOOD! Keep your paycheck! Support the Fair Tax Get into an Art museum Victorian London It's your brain -*ON WEB* - mybrainhost.com (under development) What SHOULD Government do? Much Less than it Does!

September 11th, 2006, 2:42 PM	#76 (permalink)
kavaXtreme Registered User Fresh Surpasser Joined in Dec 2005 12 posts Gave thanks: 2 Thanked 3 times	Updates For others reading this thread, you'll probably want to check out the following: Updated script (auto-delete old spam): Anyway to stop SPAM? Exim filter "tutorial" (mentioned in Bigjohn's post above): Whitelist based on subject line?

September 11th, 2006, 3:10 PM	#77 (permalink)
Bigjohn minor deity Super #1 Joined in Apr 2004 Lives in Georgia Hosted on XEON 7,340 posts Gave thanks: 23 Thanked 93 times	Yeah, you don't want to autodelete old spam... If your bayes filter gets hosed, you want to have at least 500 spam messages available to re-train it quickly. Yes, I know you'll get 500 spam's in a couple of days, but they'll all end up in your inbox! John __________________ Proud to be a Surmunity Mod! XEON PASS60 PASS61 Make a fundamental difference! My Sites: Curious about Brewing Beer? Join the community! >>>>> Some Change is GOOD! Keep your paycheck! Support the Fair Tax Get into an Art museum Victorian London It's your brain -*ON WEB* - mybrainhost.com (under development) What SHOULD Government do? Much Less than it Does!

October 19th, 2006, 11:11 AM	#80 (permalink)
grokdesigns Registered User Fresh Surpasser Joined in Oct 2006 18 posts Gave thanks: 5 Thanked 0 times	Can anybody tell me why my mail isn't being processed? I get cron job reports as follows: Code: Learning SPAM Processing /home/[name]/mail/grokdesigns.com/chris/.SPAM Learned tokens from 0 message(s) (0 message(s) examined) rm: cannot remove `/home/[name]/mail/grokdesigns.com/chris/.SPAM': Is a directory Learning HAM Processing /home/[name]/mail/grokdesigns.com/chris/.HAM Learned tokens from 0 message(s) (0 message(s) examined) rm: cannot remove `/home/[name]/mail/grokdesigns.com/chris/.HAM': Is a directory Done These are the correct directory names, but no messages are being processed...and I'm confused as to why. I have 132 messages in .SPAM and 69 in .HAM. I also don't understand why it's trying to remove the directory instead of messages. __________________ grokdesigns.com Server Name: Pass65

September 10th, 2006, 4:05 PM	#74 (permalink)
kavaXtreme Registered User Fresh Surpasser Joined in Dec 2005 12 posts Gave thanks: 2 Thanked 3 times	Bigjohn, Thanks a ton for the tutorial! I've done a lot of reading about SpamAssassin, but it's hard to find much of use for Cpanel users. Your info is the the most helpful I've found anywhere on the web. A couple of questions: 1. Could you add a section to the tutorial about which rules to modify and how to do it? 2. A significant amount of all spam I receive comes to addresses I don't use (uucp, majordomo, root, etc.). I could set up filters to discard those e-mails, but since they ONLY receive spam, they seem like ideal learning material for Bayes. Is there a way to set up Bayes to learn from them instead of using the previously mentioned method?

September 11th, 2006, 4:19 PM	#78 (permalink)
kavaXtreme Registered User Fresh Surpasser Joined in Dec 2005 12 posts Gave thanks: 2 Thanked 3 times	Sorry. I should have clarified: it doesn't auto-delete all spam, just spam that has been in the mailbox for more than the specified number of days. So you will always have a healthy dose of fresh spam in the mailbox, but spam that's been sitting around for x number of days will be deleted.

September 14th, 2006, 6:54 PM	#79 (permalink)
kavaXtreme Registered User Fresh Surpasser Joined in Dec 2005 12 posts Gave thanks: 2 Thanked 3 times	If you're like me and are not on a dedicated server, you might be interested in a reply I recently got from support: "We are sorry to say that Razor, Pyzor and DCC are not allowed." Also, you can't use custom rules (which is a major bummer for me, but I do understand there are security considerations). Still, Bayes learning and RBL checking combined with mail filters has already made a HUGE difference for me. (And I'm not even using the script with Bayes... yet.)

October 19th, 2006, 12:16 PM	#81 (permalink)
Bigjohn minor deity Super #1 Joined in Apr 2004 Lives in Georgia Hosted on XEON 7,340 posts Gave thanks: 23 Thanked 93 times	try a trailing slash? __________________ Proud to be a Surmunity Mod! XEON PASS60 PASS61 Make a fundamental difference! My Sites: Curious about Brewing Beer? Join the community! >>>>> Some Change is GOOD! Keep your paycheck! Support the Fair Tax Get into an Art museum Victorian London It's your brain -*ON WEB* - mybrainhost.com (under development) What SHOULD Government do? Much Less than it Does!

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Rate This Thread
Rate This Thread: