Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Post deleted deleted deleted deleted deleted

share|improve this question

marked as duplicate by GregD, EEAA, Shane Madden, Michael Hampton, voretaq7 Aug 7 '12 at 3:12

This question was marked as an exact duplicate of an existing question.

3  
Zero downtime is a myth. Sorry, but it's an impossibility. Look at how much money Amazon/Google/et. al. Have to throw at their infrastructure and development, and they still have downtime. – EEAA Aug 7 '12 at 2:42
1  
I should also mention...it's obvious you have a lot to learn about systems administration, networking and email services. I would strongly recommend that you start with a single email server. Run that for long enough for you to feel comfortable with its operation, backup, restore, etc. If you go forward with the system you propose, you'll quickly get underwater, and all the redundancy, security, and reliability you desire will be squandered. – EEAA Aug 7 '12 at 2:54
    
In addition to the question noted above, you should take a look at serverfault.com/questions/130300/reaching-99-9999-uptime – voretaq7 Aug 7 '12 at 3:15
2  
@FEAST email is more complicated than you think it is (As an example, you state your gmail account "hasn't dropped an email in 5 years" -- If you truly understand email you know there is no way you can make that statement with any degree of confidence). The notes on the linked questions will help get you thinking in the right direction, but reliable redundancy is a hard thing to achieve. This something you should spec and design carefully rather than "blindly attack". – voretaq7 Aug 7 '12 at 3:22
1  
@FEAST - that comment highlights how much you need to learn about email. If your server is down and someone sends you an email, it will not be lost. SMTP was designed with non-reliable networks in mind. If the sending sever is not able to deliver to the MX, it'll just queue the email and retry. Most will retry several times, up to 48 hours before sending an NDR to the sender. If your server comes up before the message timeout, your message will be delivered without issue. – EEAA Aug 7 '12 at 3:33

Avoid "reinventing the wheel"...

It's possible to obtain high-availability mail through less-complication means.

There are some redundancies already built into the SMTP protocol (retries, MX priority, etc.). You can get much of the same protection by using a backup MX that can spool if your primary delivery system is unavailable.

IMAP does what you're asking in terms of synchronization. Microsoft Exchange MAPI is a little more elegant.

The path of least resistance (and $cost) is probably obtaining an account on a hosted service, for example, hosted Microsoft Exchange available through any number of cloud providers. Let them handle the redundancy. It doesn't have to be Microsoft. There are plenty of Linux-based mail hosting solutions available, too.

See:

How to make sure that in case of our mail server being unreachable (connection down) mail still gets queued and resent once it's back up?

How to build a high availability Postfix system?

share|improve this answer
    
By spool, I mean that you can use a backup external MX relay or service to hold mail in the event that your primary mail server is down. See my edits above. – ewwhite Aug 7 '12 at 1:39

Not the answer you're looking for? Browse other questions tagged or ask your own question.