On 8/17/2017, 8:56:53 AM, Phil Stracchino <ph...@caerllewys.net> wrote: > I have a single secondary MX at a domain controlled by another competent > individual whom I know. It's useful in the event of a sustained service > outage or other delivery problems (say, if the main application server > went down and I had to rebuild it from backups).
Most sites will retry by default for 1-3 days (I think 3 days is postfix's default). A sending server will usually give a warning about a delay in the email delivery within a certain amount of time, then report failure after its configured time. If you have a backup MX, then the sender *will not know* that there is a problem. In the vast majority of cases, the perceived benefit is simply not worth the trouble. If your server is down for more than 3 days, then you have bigger problems, and the vast majority of the emails you would have held will have lost their value (if they had any real value in the first place), and the rest would have contacted the recipient by other means when they saw the delivery warnings/failures. I used to set our local warning for 4 hours (before the boss decided to migrate to Office365), because a lot of our business is time-sensitive. So, again, the actual benefit is generally far less than the perceived benefit - and there is even a real cost in many cases (sender doesn't know there is a problem), so running a backup MX, in the vast majority of cases, is simply not a good idea. But, to each their own...
