Re: [Mailman-Users] Soft bounces....
On Sun, 5 Apr 2009, Brad Knowles wrote: What do (Yahoo) mean different classes of mail? Personal, Commercial, Listmail. I can see their point. If listmail tries to send mail to a 5xx address for a couple of days, that 'lowers the reputation' of that IP more than when a personal mail gets a 5xx and the human sending it doesn't send any more mail to that address Still means that list mail has slowdowns, but personal mail won't If you're using postfix, it should be easy enough to tell it to use a given IP address when transmitting mail to yahoo.com, and use a different address when transmitting to anywhere else. This is the second postfix capability (the other was 'separate queues') that has been referenced that I find interesting. Could you kindly point me to a decent manual/doc/faq for either of these? Particularly adjusting sending IP according to sender address would be VERY cool (Or please reply personally if you think a quick reply would suffice. I don't want to clutter the list with OT material). Thanks! - Charles -- Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/archive%40jab.org Security Policy: http://wiki.list.org/x/QIA9
Re: [Mailman-Users] Soft bounces....
on 4/6/09 10:38 AM, Charles Gregory said: Personal, Commercial, Listmail. I can see their point. If listmail tries to send mail to a 5xx address for a couple of days, that 'lowers the reputation' of that IP more than when a personal mail gets a 5xx and the human sending it doesn't send any more mail to that address How do you know what is personal and what is commercial? I assume you can detect what is listmail by looking at the message or certain other factors This is the second postfix capability (the other was 'separate queues') that has been referenced that I find interesting. Could you kindly point me to a decent manual/doc/faq for either of these? Particularly adjusting sending IP according to sender address would be VERY cool (Or please reply personally if you think a quick reply would suffice. I don't want to clutter the list with OT material). Basically, you're looking to set up special postfix transport maps. A couple of examples are discussed at http://www.postfix.org/QSHAPE_README.html, but I'm sure the documentation, the FAQ, and the archives of the postfix-users mailing list would have a lot more examples that might be more useful to you. -- Brad Knowles b...@shub-internet.orgIf you like Jazz/RB guitar, check out LinkedIn Profile: my friend bigsbytracks on YouTube at http://tinyurl.com/y8kpxuhttp://preview.tinyurl.com/bigsbytracks -- Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/archive%40jab.org Security Policy: http://wiki.list.org/x/QIA9
Re: [Mailman-Users] Soft bounces....
On Fri, 3 Apr 2009, Brad Knowles wrote: Yahoo expects me to respond promptly to 5xx bounces, disabling any subscriptions, otherwise it starts generating 'greylist' 4xx bounces for my mail. Unfortunately, if I try to lower the bounce threshold, the now-daily occurence of 4xx bounces guarantees that all yahoo addresses will end up disabled. Is there any way, short of fiddling with and recompiling from source, to separate 'hard' bounce handling from 'soft' bounce handling, or to change the 'scores' that each accumulates? Thing is, Mailman doesn't see the 5xx or 4xx codes directly. It hands the messages off to your MTA, and your MTA sees those codes. Good point. Even though I'm seeing 'deferals' in the MTA logs, they don't go long enough to generate '4 hour warnings' that mailman would have to handle as a 'soft bounce'. Just verified this in the logs (where I should have looked first! LOL). Excellent! Thanks! Since we do not, and we cannot, have direct access to the 5xx versus 4xx response codes from Yahoo!, I don't think it makes a lot of sense to try to modify the code to treat a bounce resulting from a permanent 5xx response code any different from a delayed message notice resulting from a temporary 4xx response code. (nod) No, all I need to do now is lower the threshold so Yahoo won't see so many attempts to send mail to 5xx addresses Then maybe they'll stop doing so many 4xx delays of my mail (not just a mailman problem, incidentally). Set up your MTA to handle Yahoo! separately from other sites, so that any slow downs there don't slow you down in getting to other domains, and vice-versa. I use postfix, so the performance of yahoo delivery doesn't really impact other mail, other than the impact of the extra load of retries... Though I find it ironic that Yahoo sends lots of mail to me for unknown users :) Thanks again! - Charles -- Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/archive%40jab.org Security Policy: http://wiki.list.org/x/QIA9
Re: [Mailman-Users] Soft bounces....
on 4/3/09 8:50 AM, Charles Gregory said: I use postfix, so the performance of yahoo delivery doesn't really impact other mail, other than the impact of the extra load of retries... If you're using postfix, then that makes things even simpler -- by creating a separate queue for only mail to yahoo.com, you can have different queue retry rules. For example, you could retry them less frequently, and hold them for shorter periods of time, or whatever it is that the guys at yahoo want you to do. There's lots of reasons to set up separate queues for large partner domains. -- Brad Knowles b...@shub-internet.orgIf you like Jazz/RB guitar, check out LinkedIn Profile: my friend bigsbytracks on YouTube at http://tinyurl.com/y8kpxuhttp://preview.tinyurl.com/bigsbytracks -- Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/archive%40jab.org Security Policy: http://wiki.list.org/x/QIA9
[Mailman-Users] Soft bounces....
Hallo! Running into a small bit of frustration. Yahoo expects me to respond promptly to 5xx bounces, disabling any subscriptions, otherwise it starts generating 'greylist' 4xx bounces for my mail. Unfortunately, if I try to lower the bounce threshold, the now-daily occurence of 4xx bounces guarantees that all yahoo addresses will end up disabled. Is there any way, short of fiddling with and recompiling from source, to separate 'hard' bounce handling from 'soft' bounce handling, or to change the 'scores' that each accumulates? Or any other trick to handle bounces properly when yahoo is tossing out greylist deferrals. At this point is just amounts to a nuisance with delayed deliveries for a lot of list mail. But it would be better to have our mailman properly disable when 5xx errors occur - Charles, HWCN -- Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/archive%40jab.org Security Policy: http://wiki.list.org/x/QIA9
Re: [Mailman-Users] Soft bounces....
On Thu, 2 Apr 2009, Charles Gregory wrote: Yahoo expects me to respond promptly to 5xx bounces, disabling any subscriptions, otherwise it starts generating 'greylist' 4xx bounces for my mail. Unfortunately, if I try to lower the bounce threshold, the now-daily occurence of 4xx bounces guarantees that all yahoo addresses will end up This doesn't quite make sense, because 4xx isn't a bounce. It's a delay code to your mail server to try again later. sendmail, by default, generates a warning back to the sender if it hasn't been able to deliver it for 4 hours, and that notice MIGHT be interpreted as a bounce, but that's the only thing I could think of. == Chris Candreva -- ch...@westnet.com -- (914) 948-3162 WestNet Internet Services of Westchester http://www.westnet.com/ -- Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/archive%40jab.org Security Policy: http://wiki.list.org/x/QIA9
Re: [Mailman-Users] Soft bounces....
Charles Gregory wrote: Yahoo expects me to respond promptly to 5xx bounces, disabling any subscriptions, otherwise it starts generating 'greylist' 4xx bounces for my mail. Unfortunately, if I try to lower the bounce threshold, the now-daily occurence of 4xx bounces guarantees that all yahoo addresses will end up disabled. Is there any way, short of fiddling with and recompiling from source, to separate 'hard' bounce handling from 'soft' bounce handling, or to change the 'scores' that each accumulates? Or any other trick to handle bounces properly when yahoo is tossing out greylist deferrals. At this point is just amounts to a nuisance with delayed deliveries for a lot of list mail. But it would be better to have our mailman properly disable when 5xx errors occur Why is your MTA returning 4xx status to Mailman as a bounce? Your MTA should be retrying these and eventually delivering them. If you're getting bounces like message undeliverable after 4 hours; will keep trying until message is 5 days old Mailman should be ignoring those. If it isn't, send me a complete, raw copy of the DSN, and I'll fix it and tell you how to fix yours. -- Mark Sapiro m...@msapiro.netThe highway is for gamblers, San Francisco Bay Area, Californiabetter use your sense - B. Dylan -- Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/archive%40jab.org Security Policy: http://wiki.list.org/x/QIA9