Re: [Declude.JunkMail] Declude performance question
On Monday, March 21, 2005, 11:00:49 AM, Chase wrote: snip/ CS Looking at our test list (posted bellow), we likely have WAY too CS many dns blacklists. That will be the first thing I look at. Any CS other suggestions? I have had luck running a DNS server (resolver - bind) locally on the IMail box using the loopback address for the primary DNS (127.0.0.1) to speed things up. YMMV Also, take a look at this data to see which tests are performing best and perhaps weed out some of the others: http://www.sortmonster.com/MDLP/MDLP-Example-Long.html Hope this helps, _M --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type unsubscribe Declude.JunkMail. The archives can be found at http://www.mail-archive.com.
Re: [Declude.JunkMail] Declude performance question
On Monday, March 21, 2005, 11:00:49 AM, Chase wrote: CS I need some help tuning Declude for performance. Up until One other thought (pushed send too fast). You may have a test or two in there that is not responding --- causing things to time out and slow things down. If you can find it and drop it things should speed up quite a bit. _M --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type unsubscribe Declude.JunkMail. The archives can be found at http://www.mail-archive.com.
RE: [Declude.JunkMail] Declude performance question
I don' have UCEPROTECRDO, XBL-DYNA, BLKLST-SURBL andHELOISIP. Can you post your definitions for those? Can I get them off the declude website somehow (I couldn't find them)? -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc | 617.464.2440 x119 | www.bullhorn.com -Original Message-From:Pete McNeil [EMAIL PROTECTED] To: "Chase Seibert";Sent: Mar 21, 2005 11:16:28 AMSubject: Re: [Declude.JunkMail] Declude performance questionOn Monday, March 21, 2005, 11:00:49 AM, Chase wrote: CS Looking at our test list (posted bellow), we likely have WAY too CS many dns blacklists. That will be the first thing I look at. Any CS other suggestions? I have had luck running a DNS server (resolver - bind) locally on the IMail box using the loopback address for the primary DNS (127.0.0.1) to speed things up. YMMV Also, take a look at this data to see which tests are performing best and perhaps weed out some of the others: http://www.sortmonster.com/MDLP/MDLP-Example-Long.html Hope this helps, _M --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
RE: [Declude.JunkMail] Declude performance question
Any recommendations for Declude add-ins or log readers that might be able to detect the run times of various tests? If I set logging to high would it log runtimes? -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc. | 617.464.2440 x119 | www.bullhorn.com -Original Message-From:Pete McNeil [EMAIL PROTECTED] To: "Chase Seibert";Sent: Mar 21, 2005 11:28:50 AMSubject: Re: [Declude.JunkMail] Declude performance questionOn Monday, March 21, 2005, 11:00:49 AM, Chase wrote: CS I need some help tuning Declude for performance. Up until One other thought (pushed send too fast). You may have a test or two in there that is not responding --- causing things to time out and slow things down. If you can find it and drop it things should speed up quite a bit. _M --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
Re: [Declude.JunkMail] Declude performance question
The test definitions might be more helpful than the Test actions... I see AHBL is listed twice. DSbl is listed twice.Sorbs is listed as are many of the SORBS individual tests. Possible overlap. You list CBL/Blitzedall/SBL and Spamhaus listed. Potential overlap here. Reynolds is a now pay service I believe. If you aren't paying maybe that is an issue. - Original Message - From: Chase Seibert To: Declude.JunkMail@declude.com Sent: Monday, March 21, 2005 10:00 AM Subject: [Declude.JunkMail] Declude performance question I need some help tuning Declude for performance. Up until now, I have been adding tests mainly for spam filtering effectiveness. But I notice now that the server can process about7 messages per second with the spam filtering off, versus only about 1-1.5 messages per second with Declude turned on.I suspect the delays are mostly network IO bound; the processors and ram on the boxes are relatively idle. When Declude is disabled, the CPU is pinned between iMail and come custom database insertion services we have. We do about 200,000 inbound messages a day between two servers, so the speed issue really only comes up when mail has queued up for some reason and needs to be de-queued. Looking at our test list (posted bellow), we likely have WAY too many dns blacklists. That will be the first thing I look at. Any other suggestions? I don't think it's any one test, but this is what we have turned on: AHBLWARNBLITZEDALLWARNCBLWARNDSBLWARNORDBWARNSBLWARNSORBS-HTTPWARNSORBS-SOCKSWARNSORBS-MISCWARNSORBS-SMTPWARNSORBS-SPAMWARNSORBS-WEBWARNSORBS-BLOCKWARNSORBS-ZOMBIEWARNSORBS-DUHLWARNSPAMCOPWARN DSNWARNNOABUSEWARNNOPOSTMASTERWARN BADHEADERSWARNBASE64WARNCMDSPACEWARNCOMMENTSWARNHELOBOGUSWARNIPNOTINMXIGNOREMAILFROM WARNNOLEGITCONTENTIGNOREPERCENTWARNREVDNSWARNROUTINGWARNSPAMHEADERSWARN FIVETENSRCWARNspamhausWARNdsblWARNnjablWARNreynoldsWARNOridWARNJammWARNAHBLWARNsorbsWARNdnsbl-1WARNdnsbl-2WARNsurrielWARN SNIFFERWARNINV-URIBLWARN IMAILSTAT WARN -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc. | 617.464.2440 x119 | www.bullhorn.com
RE: [Declude.JunkMail] Declude performance question
Turned out to be that one of our two DNS servers was down; causing a timeout on many of the requests. That's fixed, and Declude is now performing at about 3-4 messages per second. Much better, and not that far off the throughput with zero spam filtering. -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc | 617.464.2440 x119 | www.bullhorn.com -Original Message-From: [EMAIL PROTECTED] To: "Declude.JunkMail@declude.com" <DECLUDE.JUNKMAIL@DECLUDE.COM>;Sent: Mar 21, 2005 11:32:33 AMSubject: RE: [Declude.JunkMail] Declude performance question Any recommendations for Declude add-ins or log readers that might be able to detect the run times of various tests? If I set logging to high would it log runtimes? -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc | 617.464.2440 x119 | www.bullhorn.com -Original Message-From:Pete McNeil [EMAIL PROTECTED] To: "Chase Seibert" <DECLUDE.JUNKMAIL@DECLUDE.COM>;Sent: Mar 21, 2005 11:28:50 AMSubject: Re: [Declude.JunkMail] Declude performance questionOn Monday, March 21, 2005, 11:00:49 AM, Chase wrote: CS I need some help tuning Declude for performance. Up until One other thought (pushed send too fast). You may have a test or two in there that is not responding --- causing things to time out and slow things down. If you can find it and drop it things should speed up quite a bit. _M --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
Re: [Declude.JunkMail] Declude performance question
XBL is = CBL Blitzed all UCE is probably your dnsbl-1 dnsbl-2 HELOISIP is an external test to see if the HELo is an IP address. Seehttp://www.mail-archive.com/declude.junkmail@declude.com/msg17819.html Blklst-surbl might be duplicating the invuribl test you are running. - Original Message - From: Chase Seibert To: Declude.JunkMail@declude.com Sent: Monday, March 21, 2005 10:29 AM Subject: RE: [Declude.JunkMail] Declude performance question I don' have UCEPROTECRDO, XBL-DYNA, BLKLST-SURBL andHELOISIP. Can you post your definitions for those? Can I get them off the declude website somehow (I couldn't find them)? -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc | 617.464.2440 x119 | www.bullhorn.com -Original Message-From:Pete McNeil [EMAIL PROTECTED] To: "Chase Seibert" <DECLUDE.JUNKMAIL@DECLUDE.COM>;Sent: Mar 21, 2005 11:16:28 AMSubject: Re: [Declude.JunkMail] Declude performance questionOn Monday, March 21, 2005, 11:00:49 AM, Chase wrote: CS Looking at our test list (posted bellow), we likely have WAY too CS many dns blacklists. That will be the first thing I look at. Any CS other suggestions? I have had luck running a DNS server (resolver - bind) locally on the IMail box using the loopback address for the primary DNS (127.0.0.1) to speed things up. YMMV Also, take a look at this data to see which tests are performing best and perhaps weed out some of the others: http://www.sortmonster.com/MDLP/MDLP-Example-Long.html Hope this helps, _M --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
Re: [Declude.JunkMail] Declude performance question
Test runtimes aren't listed in logging level HIGH. They might be in debug mode, but I wouldn't run that for long. - Original Message - From: Chase Seibert To: Declude.JunkMail@declude.com Sent: Monday, March 21, 2005 10:30 AM Subject: RE: [Declude.JunkMail] Declude performance question Any recommendations for Declude add-ins or log readers that might be able to detect the run times of various tests? If I set logging to high would it log runtimes? -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc. | 617.464.2440 x119 | www.bullhorn.com -Original Message-From:Pete McNeil [EMAIL PROTECTED] To: "Chase Seibert" <DECLUDE.JUNKMAIL@DECLUDE.COM>;Sent: Mar 21, 2005 11:28:50 AMSubject: Re: [Declude.JunkMail] Declude performance questionOn Monday, March 21, 2005, 11:00:49 AM, Chase wrote: CS I need some help tuning Declude for performance. Up until One other thought (pushed send too fast). You may have a test or two in there that is not responding --- causing things to time out and slow things down. If you can find it and drop it things should speed up quite a bit. _M --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
RE: [Declude.JunkMail] Declude performance question
Here is an idea many of us use: Configure DNS service on the Imail server for caching only. Point Imail/Declude to use the DNS on the same server. Configure the DNS service to use your other DNS servers and forwarders. You will see a slight performance boost. John Tolmachoff Engineer/Consultant/Owner eServices For You -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Chase Seibert Sent: Monday, March 21, 2005 8:46 AM To: Declude.JunkMail@declude.com Subject: RE: [Declude.JunkMail] Declude performance question Turned out to be that one of our two DNS servers was down; causing a timeout on many of the requests. That's fixed, and Declude is now performing at about 3-4 messages per second. Much better, and not that far off the throughput with zero spam filtering. -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc | 617.464.2440 x119 | www.bullhorn.com -Original Message- From: [EMAIL PROTECTED] To: Declude.JunkMail@declude.com <DECLUDE.JUNKMAIL@DECLUDE.COM>; Sent: Mar 21, 2005 11:32:33 AM Subject: RE: [Declude.JunkMail] Declude performance question Any recommendations for Declude add-ins or log readers that might be able to detect the run times of various tests? If I set logging to high would it log runtimes? -Chase Chase Seibert| Network and Systems Engineer | Bullhorn Inc | 617.464.2440 x119 | www.bullhorn.com -Original Message- From:Pete McNeil [EMAIL PROTECTED] To: Chase Seibert <DECLUDE.JUNKMAIL@DECLUDE.COM>; Sent: Mar 21, 2005 11:28:50 AM Subject: Re: [Declude.JunkMail] Declude performance question On Monday, March 21, 2005, 11:00:49 AM, Chase wrote: CS I need some help tuning Declude for performance. Up until One other thought (pushed send too fast). You may have a test or two in there that is not responding --- causing things to time out and slow things down. If you can find it and drop it things should speed up quite a bit. _M --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type unsubscribe Declude.JunkMail. The archives can be found at http://www.mail-archive.com.
Re: [Declude.JunkMail] Declude performance question
Chase Seibert wrote: I don' have UCEPROTECRDO, XBL-DYNA, BLKLST-SURBL and HELOISIP. Can you post your definitions for those? Can I get them off the declude website somehow (I couldn't find them)? I wrote the HELOISIP test. It's not a DNS test. Rather, it will tell you if the HELO or REVDNS of the incoming message has an embedded IP address. While not 100%, this is often an indicator of a spam message from a dynamic IP address. You can download it from http://bud.thedurlands.com Hope this helps -- I don't need anger management; I need people to stop making me angry Bud Durland, CNE Mold-Rite Plastics Network Administrator http://www.mrpcap.com --- [This E-mail scanned for viruses by Declude Virus / Sophos AV] --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type unsubscribe Declude.JunkMail. The archives can be found at http://www.mail-archive.com.