Does anyone have any good tricks for implementing a redundant POP3 server? I was thinking something like this:
Put primary server on 10.0.0.1 and put secondary server on 10.0.0.2. Use DST-NAT rule to direct public address of server to 10.0.0.1. Use netwatch to monitor 10.0.0.1 and on failure use a script to change the DST-NAT rule to 10.0.0.2.
The problems I see with this are:
Primary server may still respond to netwatch when mail server software is frozen or otherwise broken. So some test of port 110 would be needed?
Manual transfer of message queues from secondary to primary server will be required to switch back to primary.
Messages queued on primary server will be lost if the failure is catostrophic.
POP Server1: x.x.x.2
POP Server2: x.x.x.3
VRRP Addr: x.x.x.1
MT Routes to x.x.x.1, if 2 is down, VRRP maps 1 to 3, if 3 is down, VRRP maps 1 to 2. How to implement VRRP on your systems, is beyond this forum however.
There are loads of manners to create redundancy between servers - very few of them have anything to do with routers… Google a bit.
VRRP appears to require the servers to be VRRP compliant, or introduce and additional router in front of each server, so that’s not ideal from a cost and complexity view point.
I’ve googled long and hard and the only type of setup I can find is where you have two of more servers all using a common storage medium. This is fine except that the storage becomes the point of failure. I guess big enterprise use raid arrays and other tricks to give the NAS or SAN very high availibility. I am looking for something more ‘low end’
Can you name some of the other methods of implementing server redundancy so that I can google for those terms?
Your point of storage is not a single point of failure. Read up on SANs (Storage Area Networks). Again, this is way of topic here, and appart from that anything (even a switch) can be made redundant.
It’s just a question of how much money you are willing to spend on it
