Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

We recently had a harddisk failure on our mail server (debian/postfix/courier) so I'm researching a way to build redundancy.

The first way I found was to setup multiple mx Records in the DNS Server so that the second Mailserver jumps in when the first fails. While this is easy to build it's not what I have in mind for redundancy.

The second one is to use a third box which provides a NFS share where the Mailboxes are stored. This isn't really an option, because it provides a single point of failure like a standalone Mailserver.

I would like to have a way to synchronize the two servers, so that, if the first Mailserver fails, the enduser doesn't notice there had been a failure. (Mails still in the IMAP Mailbox etc.)

Any chance?

share|improve this question

1 Answer 1

up vote 1 down vote accepted

You should setup a disk redundancy system. You have two immediate solutions :

  1. You use RAID 1 or RAID 10 to mirror your disks
  2. You have now multiple mail server so you can use DRBD to synchronize the disks over the network. -> http://www.drbd.org/home/mirroring/
share|improve this answer
    
Thank you for your reply. The failing Mailserver had a RAID 1 configuration. Strangely the S-ATA controller seemed to fail. It only recovered after a complete power down (reset didn't help). I will definitely look into the DRBD solution. –  Christoph Winkler Aug 29 '11 at 7:09
    
Expect to run into problems trying to use DRBD for replication - particularly for anything other than maildir/mailbox storage. –  symcbean Aug 29 '11 at 11:36
1  
I indeed start from the postulate that he will not try to sync the operating system but just the storage partition ! –  DrGkill Aug 29 '11 at 15:22

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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