A bit of background on our architecture, currently have two active directory sites and two DAGs stretched to opposing AD sites. The DCs are highly connected by redundant 10Gb circuits. Running exhange 2010 sp2 ru4 in prod and I have been testing the below in lab after upgrading to 2010 SP3 (still on server 2008R2)
I have noticed a couple of bugs that I won't go into too much detail in but it revolves around outlook anywhere and accessing shared mailboxes over HTTPS encapsulation. Also sent items location when set to "senderandfrom" doesnt work too well when mailboxes and shared mailboxes span clusters or DAGs.
With OutlookWebApp we set redirecttoOptimalOWAserver to $false - this works well for us as we have flexibility to make a DNS change with a short TTL and mailboxes can be accessed via any CAS server over IIS..
I came across this -
I set the registry value AND I set the AllowcrosssiteRPCaccess on the DAG to $true, when I mount a DB in the opposing DC, it does get rid of the annoying "your exchange administrator has made changes that requires you to restart" and it does continue working BUT ONLY throught the original client access server. If I change the RPC client access server/array in the local hosts file to the CAS array in the site where the mailbox is mounted it does not work. I need to restart outlook / get autodiscover new MAPI/RPC endpoint and then it works.
It really doesnt matter from a client experience perspective whether they connect to CAS array A or CAS array B. Is what I'm trying to do achievable? The reason I'd like to do this is so that we can control via DNS where clients connect to. The benefits I see:
1. No having to restart outlook clients after mailbox moves.
2.fix the bug where sent items don't go to the shared mailbox sent items when the shared mailbox is in a different AD site
Has anyone actually tested EnablePreferredSiteEnforcement?