Link to home
Start Free TrialLog in
Avatar of dmurawski
dmurawski

asked on

SMTP Connectors in same Exchange Org.

I have two smtp connectors for outbound mail in the same org, the connectors are in different admin groups with cost 1 and cost2 for the entire org.
I wanted the second connector to be used if connector 1 was unavailable.   During a recent maintenance window I was able to test this config and it was not successful from 3 sites.
The outbound mail was stuck in the queue on 3 sites, on 2 sites mail did flow out the second smtp connector.

Site 1  1 backend server  
Site 2  1 backend server -  1 frontend server smtp connector cost 2
Site 3  1 backend server
Site 4  2 backend servers - 1 frontend server smtp connector cost 1

I took smtp connector cost 1 offline.    On Site 1 and Site 2 mail did route through successfully through the smtp connector on site 2. On site 3 and site 4 mail queued and did not route mail queue on both until I brought the smtp connector cost 1 online.

Any input on this would be great why this didn't work on site 3 and site 4.

Thanks.
Avatar of MATTHEW_L
MATTHEW_L

The connectors are in different admin groups doesn't really matter too much, admin groups are just for assigning rights etc and grouping servers.  What about routing groups?  If they are in different routing groups do you have routing group connectors?
Do you mean admin groups or routing groups?

Admin groups play no part in routing of email - they are admin boundaries.

Exchange uses the first SMTP Connector that it comes to that fits its requirements. If the connector is offline (ie still exists but cannot be sued) then Exchange will simply try and queue the email. It will not try and use a different connector. If you are looking to do some kind of failover then I am afraid that isn't possible using SMTP Connectors.

As such the behaviour you have seen is what I would expect.

Simon.
Avatar of dmurawski

ASKER

I understand that the admin groups are just admin boundardies.
The routing groups for each are configured this way.

Example.

Admin Group (Site 1)
Routing Group (Site 1)
Site 1 to Site 2 Routing Group Connector
Site 1 to Site 3 Routing Group Connector
Site 1 to Site 4 Routing Group Connector

Site 2 to Site 1 RGC
Site 2 to Site 3 RGC
Site 2 to Site 4 RGC
SMTP Connector cost 2.

Same for Sites 3 and 4.  Site 4 has an SMTP Connector with cost 1.

So what you are telling me is that you can't have multiple SMTP Connectors for failing over purposes?   So if Site 4 SMTP Connector with Cost 1 goes down,  the SMTP connector with Cost 2 in Site 2 won't route outbound mail?

ASKER CERTIFIED SOLUTION
Avatar of icky2000
icky2000

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
But if the connector itself goes down or the server hosting the connector, this will not help.
The SMTP Connector is not configured to use smart hosts.  The 2 front-end servers are the only servers allowed to send outbound smtp.

Lets be clear: SMTP Connectors do not and cannot go down. They are not a physical thing tied to a physical resource. They are a logical thing that exist at the org level. If you decide to assign only one SMTP Virtual Server to an SMTP Connector, of course that becomes a single point of failure. You shouldn't do that. You should use at least two and they should be on separate boxes.
Here is the way the SMTP Connectors are setup.

(Site 4) The SMTP connector (cost1) is using DNS to route.  There is a local bridgehead server listed for the default SMTP Virtual Server.

(Site 2) The SMTP connector (cost2) is using DNS to route. There  is a local bridgehead server listed for the default SMTP Virtual Server.

If the bridgehead server is down in Site 4 goes down, then the smtp connector (cost1) won't route outbound mail.  I thought that the second SMTP connector would be able to route the mail out.