Hello all!

I’m planing to setup a mx backup with Mailcow (for a primary Mailcow mx). I have some questions and doubts… If you could help me see things a bit clearer, that would be great.

Mx backup spec
I don’t think the server supporting this secondary service needs to be very powerful: no SOGo and “minimal” storage (it will only relay).
Do you think my assumptions are right, or am I missing something? What do I need to think about when sizing this server?

Domains creation
The creation of domains on the primary server is done in self-service mode by my customers. Now, the relayed domain will have to be created each time on the secondary server too.
What advice would you give me to automate this procedure? On the primary server, is there a “trigger” (at creation) that I could use to make an api call to the secondary? Or, do you think this “orchestration” should be carried out externally?

It’s possible (although I’ve read it carefully) that I’ve missed such information in the documentation. If so, please let me know.

Thanks in advance for all your tips
Pierre

Seems like your requirements engineering is not yet done.
What you want do cover? HA? DR? What RPO, what RTO?
What is the business continuity plan?

Have something to say?

Join the community by quickly registering to participate in this discussion. We'd like to see you joining our great moo-community!

My only need is “inbound” HA, definitely not DR - my DRP relies on cold-standby, snapshots and Ansible scripts and that’s a different topic than mx backup.
I want to make sure that:

  • If my primary mx server (which also supports SOGo) is no longer accessible, incoming mails won’t be lost. - in that case, I understand very well that all other services will be down.
  • As soon as the primary server is up again, I can process incoming mail queues

Thank you to ask me to clarify my question.

Mails get only lost when your mailcow is not available for more than 48 hours usually. Then the sending mailservers will start sending NDRs to the senders.
And even then they are not “lost”, the sender knows that it did not went through (he is informed after 4 hours about a possible delayed delivery).
Actually you would only need a secondary postfix which hoards the incoming mails and then once the mailcow is back up running relays them to mailcow.

OK. Thank you for your help & your time.

No one is typing