All, We have a pair of LTMs load balancing traffic for all Exchange clients using one Virtual Server. Currently, we're having an issue where the OWA interface (through IE 8) for Exchange 2010 will hang when traversing the F5, and also encounters a large number of errors with javascripts used by OWA. If any of the specific servers in the group are used directly, OWA behaves as expected. I've seen other comments here and most seem to point to some kind of issue with compression settings in the HTTP profile used on the VS but I haven't seen anything definitive. Has anyone found out what causes this, and better yet, specifically how to resolve the issue? Thanks in advance. Jerry

Jerry, I had a similar issue. By unchecking "Keep Accept Encoding" from the http profile, errors with javascripts disappeared. OWA no longer will hang. Jamie

I thought I read, though, that doing this will break other functionality (like web services??). Is that not the case?

I read that too. We are not using ActiveSync so I didn't worry about it. I have a separate VIP for OWA, RPC and Autodiscover. I only remove "Keep Accept Encoding" for OWA. I don't know if there a better way, but we had no issue for over a month now.

Yah... I have them all tied to the same VIP. Our ActiveSync clients aren't internal but I definitely don't want to inadvertently break anything with RPC, Autodiscover, or Outlook Anywhere. Thanks for the feedback, though. Much appreciated. Hopefully, someone from F5 reading this might post something back to provide more insight into this.

Is compression enabled on the IIS?

Exchange 2010 OWA Hangs

25 Replies

jdewing
Cirrus
May 27, 2011
Jerry,

I had a similar issue. By unchecking "Keep Accept Encoding" from the http profile, errors with javascripts disappeared. OWA no longer will hang.

Jamie
Gerald_G__Young
Nimbostratus
May 27, 2011
I thought I read, though, that doing this will break other functionality (like web services??). Is that not the case?
jdewing
Cirrus
May 27, 2011
I read that too. We are not using ActiveSync so I didn't worry about it. I have a separate VIP for OWA, RPC and Autodiscover. I only remove "Keep Accept Encoding" for OWA. I don't know if there a better way, but we had no issue for over a month now.
Gerald_G__Young
Nimbostratus
May 27, 2011
Yah... I have them all tied to the same VIP. Our ActiveSync clients aren't internal but I definitely don't want to inadvertently break anything with RPC, Autodiscover, or Outlook Anywhere.

Thanks for the feedback, though. Much appreciated. Hopefully, someone from F5 reading this might post something back to provide more insight into this.
Michael_Koyfma1
Cirrus
May 27, 2011
Is compression enabled on the IIS?
Gerald_G__Young
Nimbostratus
May 27, 2011
Why, yes, it is; appears to be a default. Do you think the double compression is the issue here?
MJS_62463
Nimbostratus
May 30, 2011
Can I ask how you have all setup your Exchange Virtual Directory Internal/External Url's? (Sorry to hijack this thread) Are your InternalUrls pointing to the individual servers? And have you left the External Url's empty? Or do they point to the F5('s)? As I understand it, in a multiple CAS environment the OWA External Url must be empty in order for Proxying to work across sites. Anyhow, I would love to to know what the proper configuration is as I cannot find documentation explaining it fully.
Gerald_G__Young
Nimbostratus
Jun 13, 2011
MJS,

The internal URLs point to the specific servers and the external URLs point to our load balanced FQDN, which the F5 is servicing. We only have one AD Site, however.

What's interesting about this is that if we access OWA directly via one of the CAS servers, subsequent access via the load balanced FQDN on the F5 works fine.

Since no one else has responded to this thread, I'm going to go ahead and open up a case with F5 Support. I'll post whatever findings we uncover during that support call.
MJS_62463
Nimbostratus
Jun 13, 2011
One thing I can add is...in our case I have left all external URL's empty. Apparently this is necessary in order to get OWA proxying to work. Everything seems to be working pretty well for us now although I am still not convinced this configuration makes proper use of the load balancer(s).
Gerald_G__Young
Nimbostratus
Jun 13, 2011
That didn't make a difference for me, although I suppose I might need to reboot the CAS servers after having made that change. At TechEd this year, they did make mention of SP2 for Exchange 2010 fixing the issue with OWA being redirected to another set of CAS servers and requiring the user to authenticate a second time. That doesn't sound like what you're referring to, though.