Repeating messages stating that the "Connection to the Exchange server has been lost". (96422)

The following Messages repeting in the Collaboration Services Events.

![2012/08/02 19:01:08.802] (2) Warning :<Exchange Layer> (32623) Connection to the Exchange server has been lost.(Network error or the Exchange server is down.)
![2012/08/02 19:11:25.893] (2) Information :<Exchange Layer> (32624) Connection to Exchange server has been restored.
![2012/08/02 19:28:26.491] (2) Warning :<Exchange Layer> (32623) Connection to the Exchange server has been lost.(Network error or the Exchange server is down.)
![2012/08/02 19:39:07.397] (2) Information :<Exchange Layer> (32624) Connection to Exchange server has been restored.

Also Syncing large publications will not complete.

When looking at a packet capture taken during a Connection Lost event you will see "Local limit exceeded" messages sent from Exchange Server defined as the connection server.

Example from Packet Capture where 192.168.0.7 is the QCS server and 192.168.0.1 is the Exchange Server QCS is configured to connect to:

192.168.0.7 192.168.0.1 DCERPC Bind: call_id: 2 Fragment: Single, 2 context items, 1st NSPI V56.0
192.168.0.1 192.168.0.7 DCERPC Bind_ack: call_id: 2 Fragment: Single Unknown result (3), reason: Local limit exceeded
192.168.0.1 192.168.0.7 DCERPC Fault: call_id: 2 Fragment: Single ctx_id: 0 status: nca_s_fault_access_denied
192.168.0.1 192.168.0.7 TCP 59533 > 55331 [FIN, ACK] Seq=117 Ack=201 Win=131072 Len=0 2012-08-09 15:15:48.131472

Or (192.168.86.185 being QCS, 192.168.80.31 being Exchange server QCS configured to connect to): 

Or when testing the service account mailbox you may get the following errors Error "0X80040111" or  "0X80040115"

This can be caused by a Hardware Load Balancer.

WORKAROUND:

Bypass the Hardware Load Balancer.

1. From within the Collaboration Services Console, right click Collaboration Services and select Properties.

2. On the General Page for "Exchange Server" enter the host name of a specific Client Access Server (CAS) and not the name of the Array or Load Balancer.

**Note: In some cases requests to connect to the host name of a specific CAS Server may be re directed back to the Load Balancer.

In this case you will need to add an Entry to the local Hosts file on the Collaboration Services server, so that Collaboration Services will connect directly to the Exchange Server and Bypass the Hardware Load Balancer.

3. Add an entry to the Hosts file with the name of the Array or Load Balancer and the IP address of the specific CAS server.

You should add two entries, the FQDN of the Load Balencer as well as the NETBIOS name. 

E.g If the Physical IP address of the CAS Server is 192.168.0.50 then the host file would look like this

      192.168.0.50 mail

      192.168.0.50 mail.company.local

4. Save the hosts file, run ipconfig /flushdns 

5. Verify the Exchange Server has been lost error's are no longer present.