On Tue, Jun 16, 2015 at 8:59 AM, Steve Matzura <numb...@noisynotes.com> wrote:
> I have finished setting everything up on a test system using a > different flavor of Linux and a more current version of everything > than my production system. Let's call them prod.example.com and > test.example.com. Without interrupting mail service on prod, which is > half of what that system does so I really can't take it down and wait > for DNS changeovers back and forth, what can be done on test to make > it look like and work like prod? For instance, when I start Postfix on > test, it's trying to deliver messages to prod and is unable to. I > could extract stuff from maillog which might be of some help to figure > out what's going on, but before I do that, is it even possible to do > what I'm wanting to do--spoof my current Dovecot+Postfix setup to > think it's on prod when it really isn't? By the way, it's OK for > messages from test to get into prod, people on the mailing lists on > prod know this could and probably will happen. > So... I guess "prod" has the mailboxes, and you want to test " test.example.com" as a prospect replacement for prod. If that's the case, you will want to enable all the corresponding "local" delivery in test, and furthermore it could even start thinking it is "prod" (even though it will still only respond to its address for "test"). After doing this, you can configure an account on your mail client to connect to "test", and do most of the tests there. This scenario is pretty common when you are configuring a new system, so, indeed is possible, and there are several ways of doing it, depending on the details of what you want to do. You could even setup a test subdomain in order to do a complete test including external mail sending, and before promoting to production. Now, the switchover planning (or "promoting test as prod") is another history, and can be done by several different means, one of those being using (or creating and then using) a private network and redirecting traffic on prod to test system, and then doing the DNS change, effectively making all traffic that would originally go to prod, go to "test" (that now would be called prod, but I need a way to distinguish them), ... doing this would either expose you to some spam going through or require some heavy usage of advanced routing, so, before doing this it is recommended to have DNS TTL set to something like 60 seconds or so. After 2 minutes has passed, all new connections should be going to your new prod, and you should be able to stop prod. Oh, but there is more: what about mailboxes? (likely maildirs) that's yet another point that require planing, and will depend on your mailboxes format, so, won't start with that right now. Well, I hope this is useful, and if you want more help, please elaborate a bit more on what you want to do. Sincerely, -- Ildefonso Camargo Command Prompt, Inc. - http://www.commandprompt.com/ PostgreSQL Support, Training, Professional Services and Development High Availability, Oracle Conversion, Postgres-XC @cmdpromptinc - 509-416-6579
