Single Exchange 2007 Server – Outlook Anywhere (RPC/HTTPS) Performance Issue Fix

Here’s the inventory:

  • Single Exchange Server (2007) on Window Server 2008
  • 100 Users, all connecting via RPC/HTTPS (Outlook anywhere or Webmail)
  • Domain Controller (1 of 2) is on the Exchange Server itself

 

The problem:

Although I’m not positive on the timing of the issue, it would appear we had Exchange Server SP1 installed. However, once hotfix 6 was applied (and it did not go well the first, time you must ensure you are in the Exchange Administrators group, domain admins isn’t enough, it can’t see the CAS role and breaks OWA), it seemed Exchange was somewhat ‘reversed’ or broken.

 

The issue is, it was resulting in the server sometimes becoming totally unresponsive (to RDP, very slow OWA refresh, etc.) and Outlook Anywhere users seeing a constant “send/receive” that would never finish. The only way to overcome this was to reboot the server, and it was starting to approach a twice a day activity.

 

The fix:

Here’s what we discovered and the references. The problem is, Exchange configured to provide Outlook Anywhere on the same machine as the mailbox server (and let’s face it, sometimes budget comes before implementing lots of servers to divide roles…) needs help with knowing where/what LocalHost is. And on a Window Server 2008 box where IPv6 (really, why? Is it time to care yet really?) is, Localhost becomes even more difficult for Exchange to find.

 

The fix, in the end was simple:

  • Disable IPv6 in your Network Stack
  • Modify your local hosts file with the following:
    • Remark out the ::1 localhost entry in hosts (use a #)
    • Add entries for your local server name (hostname and FQDN) and any other SAN names you’ll have there for your UCC cert

Hope this helps someone. I understand this issue is being addressed properly for Exchange Server 2007 Service Pack 2, but until then…

 

References: