Hub Transport Redundancy (Full Version)

All Forums >> [Microsoft Exchange 2007] >> High Availability



Message

rparsons1000 -> Hub Transport Redundancy (20.Jun.2007 10:08:25 PM)

We had some sort of failure today on one of our Hub Transport servers, the first one ever brought up in a CCR environment. It was giving errors about journaling not working, then Microsoft Exchange couldn't retrieve the new Transport server configuration. We have 2 Hub Transport servers in this AD site but the other one never picked up to send and deliver email to and from an Exchange 2003 connector. The mail queued up for an hour before I heard about the problem and restarted the HTS.
 
How do you provide redundacy for Hub Transport servers when still using an Exchage 2003 connector to get to and from the internet as well as 2003 users? I thought this would be transparent and can't seem to find any information.


 
 





Henrik Walther -> RE: Hub Transport Redundancy (24.Jun.2007 5:54:02 AM)

In that situation you might want to create an additional RGC between the E2K3 bridgehead and the second HT server.

More info on how to create RGCs between the legacy and E2K7 system, see:
http://technet.microsoft.com/en-us/library/aa997292.aspx



BeTaCam -> RE: Hub Transport Redundancy (25.Jun.2007 12:59:32 AM)

Actually, you can do an NLBS of the hub transport servers. On the Virtual interface just ensure that a listener on port 25 is excluded. (removed).
 
it works perfectly fine and i have not seen an MS article that explicitly calls out saying not supported. probably they dont know that it can be done, thats all.
 
Hth
 
BC
 
coca cola cans are red in india too. !!!



Henrik Walther -> RE: Hub Transport Redundancy (25.Jun.2007 2:55:25 PM)

But they don't mention anewhere this is supported, so you should definately have this solution verified by MS before using it in a corporate production environment!



rparsons1000 -> RE: Hub Transport Redundancy (25.Jun.2007 10:36:51 PM)

I like the idea of the second connector until 2003 is gone. Unfotunately my CAS roles are on the same servers due to finances. If only MS Virtual Server supported 64 bit hosts [&:]

With this said, would NLB the way to go (port 80 & 443) and still create the second connector? Will I have to uninstall prior to setting up NLB?

I would love to just dump CAS but correct me if I am wrong, I believe autodiscover, free/busy and Public Folders with Exchange 2003 still in the network  require CAS role(s).



Henrik Walther -> RE: Hub Transport Redundancy (10.Jul.2007 2:26:30 AM)

Replied is a little late, but yes can configure the NLB without removing the E2K7 server roles first.



rparsons1000 -> RE: Hub Transport Redundancy (10.Jul.2007 9:09:46 AM)

Thanks Henrik. I did my first NLB on some web servers last week and it was very easy. I understand how this is done now and will get it done prior to production (waiting for SP1).



rparsons1000 -> RE: Hub Transport Redundancy (26.Jul.2007 10:39:02 PM)

We had another failover to the second Hub server so was forced to get this fixed. I created a second connector but the queued mail just sat so I did some looking and found that the original connector created during install only had the first Hub in as a bridgehead. Used the set-routinggroupconnector cmdlet to put the second Hub server on both connecotrs (2003 & 2007 sides), restarted and voila, all mail went out from the second Hub.

I still need to NLB for the CAS role but expect to have an extremely resilient topology when complete.



Henrik Walther -> RE: Hub Transport Redundancy (28.Jul.2007 6:17:24 PM)

Well the timing is good then, because MS won't supported CAS NLBs before SP1 RTWs.



Page: [1]