> -----Original Message----- > From: cctalk [mailto:[email protected]] On Behalf Of Jay West > Sent: 02 July 2016 10:37 > To: 'General Discussion: On-Topic and Off-Topic Posts' > <[email protected]> > Subject: RE: Excessive Bounces > > I did dig into this for Rob, as well as a smattering of others that got disabled > yesterday early am. > > Almost 100% of the email addresses that get disabled due to excessive > retries are Hotmail, aol, and yahoo. Those email providers are well known as > "less than reliable" to many servers, often because they are major sources of > spam. > > Of those bounces, almost 100% of them were due to DMARC/DKIM > mechanisms. I view DMARC/DKIM as less than useless. First, they are far > from universally implemented - so most valid email is not DMARC/DKIM > "protected". Second a large percentage of the spam in the world *IS* > DMARC/DKIM "protected". > > What is odd to me, many of the email addresses (list members) that get > dropped due to excessive retries are listmembers who *DO* get list traffic > for a while before getting occasionally disabled. I would bet that what is > happening most of the time is that their ISP maintains a flock of MX hosts and > at one time decided to implement DKIM/DMARC but then realized it's a bad > idea and removed it from service but forgot to do that on one or two of the > flock. Thus... by happenstance once in a while your email hits that particular > mx host and gets toasted. The reverse is certainly possible - that they are > "beta testing" it on a few of their MX hosts but not wanting to do it > everywhere. > > Please, let's not go into a long thread of the merits of SPF/DKIM/DMARC... > I'm just letting folks know what is probably responsible for most of the > "subscription disabled..." stuff. >
I have no competence to judge the merits or otherwise of these mechanisms, so I won't say anything about it. What I would say though is, if this affects a fair number of people, could the list server be configured to be a bit more tolerant of bounces? Regards Rob
