What happens to mailflow during maintenance mode?

What happens to mailflow during maintenance mode?

There is a function in Exchange MP that test mailflow between mailservers. A friend asked me what happens to those test if one mail server is in maintenance mode?

I have two mail servers, MX01 and MX02. I run the wizard and setup mailflow test between them. After a while I start receive notifications every 10 minute that MX02 has sent a e-mail and MX01 received it.

I put MX01 into maintenance mode and waits 5 minutes to make sure that the maintenance mode flag has been changed in the OnePoint database, then I shut down MX01.

During the time when MX01 is in maintenance mode there still is events that MX02 has sent e-mails and all local test is ok. There is no alerts.

30 minutes later when MX01 returns from maintenance mode I received two alerts about missing heartbeat, that is normal because the server is still down.

10 minutes later I start MX01 again, all normal events under events in operator console. Mailflow events start coming back again, both that MX02 sends and MX01 receives. 10 minutes later this error is generated from MX01

Computer: MX01
Description:
Specified delivery latency of mail flow messages exceeded. Detailed information:
Delivery latency from MX02 (at Fri Jul 28 12:19:01 UTC+0200 2006) = 2926 (seconds)
Delivery latency from MX02 (at Fri Jul 28 12:49:00 UTC+0200 2006) = 1127 (seconds)
Delivery latency from MX02 (at Fri Jul 28 12:34:00 UTC+0200 2006) = 2027 (seconds)
Delivery latency from MX02 (at Fri Jul 28 13:04:00 UTC+0200 2006) = 227 (seconds)
This event was generated by the script: "Exchange 2003 - Mail flow receiver"

2000 seconds is around 30 minutes, the same as the maintenance mode.  

Summary: There is no alert generated when one of the mail servers is in maintenance mode.  At least not the first 30 minutes.

About

Microsoft

One thought on “What happens to mailflow during maintenance mode?

  1. I tried this out in an enviroment with 42 exchange servers and can confirm that it was working as expected even in that case.

    Can you find out if there is anyway to get raid of client side monitoring when the traget is put in maintenance mode?

    Nice post.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.