I've run into this problem at more than one high volume postfix site, on different versions of RHEL/CentOS/Fedora, and different versions of postfix (all of them 2.2 or higher), and the one common factor is yahoo.
So I wonder if any of you administer high volume postfix sites and have run into the same problem, and if you've found any workarounds. What happens is that messages to yahoo.com build up in the active queue, and in some cases fill it up (even if we increase the active queue by a lot, if the mailing is big enough). Messages to yahoo are delivered more slowly than messages to gmail, hotmail, comcast, and other prominent ISPs, and it's always yahoo that clogs up the active queue. Usually, messages to yahoo are still being delivered hours after the rest of a large mailing has completed (except deferred messages). One symptom is that yahoo messages are generally not deferred, so in a setup with multiple first-pass MTAs and a fallback relay, they yahoo messages build up on the first-pass MTAs, not the fallback relay. Yes, on some occasions yahoo seems to be deliberately rate-limiting, issuing 421 or 451 deferrals with codes to check on their web site. I've dealt with that several ways, but that's a different issue. This problem I'm describing happens even when 90%+ of postfix log entries with the string "yahoo.com" are of the status=sent variety. I've tried futzing with the connection cache reuse limit, the smtpd concurrency limit, active queue sizes, and some other things, but haven't found a good solution yet. It's always Yahoo. And, as I say, I've seen this at more than one site. Any ideas? -- Cos
