Scenario: After the installation of additional Exchange 2013 servers, we noticed that Outlook Anywhere is broke in our Exchange 2010 environment. The Exchange 2013 environment and mailboxes are unaffected by the connection problems, but the Exchange 2010 mailboxes cannot use Outlook Anywhere. Ex2010 mailboxes have to make a connection directly to the CAS Array or connect via other protocols as workarounds.
Symptoms:
◦We are seeing 503 HTTP Statuses (Service Unavailable) in the Ex2013 IIS logs when trying to connect to Exchange 2010 for Outlook Anywhere.
◦Outlook Clients either cannot establish a Outlook Anywhere connection, or their connection failback’s to a standard TCP connection.
◦Exchange 2013 mailboxes may have problems connecting to Public Folders on Exchange 2010 via Outlook.
◦Running the powershell command below fails when trying to get a referral and gives a 0x0000006BA or 0x6Ba error:
test-outlookconnectivity -protocols HTTP -credential $mycreds -verbose
◦Using the following command, you are unable to make a connection to ports 6001, 6002, and 6004.
rpcping -t ncacn_http -o RpcProxy=mail.domain.com -P “user,domain,password” -H 1 -F 3 -a connect -u 9 -v 3 -s casarray.domain.com -I “user,domain,password” -e 6001
Cause: Ex2010 has a limit of servers that can be entered into a ServerFarm value in the Registry. The addition of the new Ex2013 servers took us over the 64 server limit. When this happened, Outlook Anywhere (RPC/HTTP) in Ex2010 broke. Troubleshooting various symptoms pointed to different areas of Exchange that could be the cause of this behavior. Later it was discovered that it was a bug confirmed by Microsoft. The real kicker of this ‘bug’ is that the Exchange 2013 servers, regardless if their frontend (CAS) or backend (MBX) servers, gets populated in this Ex 2010 key because of the Ex2013 architectural differences.
2 Resolutions:
1. Microsoft has an IU (interium update) for Exchange 2010 to fix this issue. You just need to remove this IU before you proceed with other installs. Supposedly this issue will be fixed in Exchange 2010 RU9 for SP3. This was our fix
Or
2. You can edit the registry manually. Note this worked for us for a while, but we did have problems with this and resorted to the IU method. Regardless , this method may get you by for a little while:
Turn the Polling for the RPCHttpConfigurator off by going setting the PeriodicPollingMinutes to 0:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\services\MSExchangeServiceHost\RpcHttpConfigurator\PeriodicPollingMinutes
Manually remove the Exchange 2013 servers from the ServerFarm Registry Key to take the number of entries below 64:
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Rpc\RpcProxy\LBSConfiguration\ca5b08e5-4a52-5701-0000-000000000000