Exchange 2010 OWA Hangs

Gerald,

 

 

Just trying to understand your setup a bit better. You have externalURL specified to be F5's VIP - is the VIP available internally and externally? Typically, you want to have all URLs on CAS to be setup identically - if internal and external are the same, please set both URLs to be the same. Did you check the compression issue to make sure it's not causing a problem here?

 

 

Michael,

 

 

We use the same FQDN for both internal and external access. Externally, the record points to a VIP that runs on a pair of TMG servers, which then load balance between the servers directly as opposed to passing through the LTMs. Internally, the record points to a VIP that runs on a pair of LTMs. Are you saying that if the record is myexchangeservices.mydomain.com that the URLs configured on the CAS servers for all web services need to be configured to point to https://myexchangeservices.mydomain.com/ (owa, ecp, Microsoft-Server-ActiveSync, oab) for both the internal and external URLs?

 

 

Disabling compression at the IIS level didn't resolve the issue.

So, you're using split-horizon DNS - that's good. I still recommend defining both Internal and External URLs on CAS servers, even if they are the same. What version of F5 are you running? Have you configured all the Exchange-related stuff according to the deployment guide? Are you doing the SSL offload on F5?

 

 

We're using BIG-IP 10.2.0 Build 1707.0 Final. We configured according to the deployment guide (we created the Virtual Servers prior to upgrading to 10.x, so didn't use the template) but had to remove Nagle's Algorhithm as that caused significant connection delays with Outlook. SSL connections are terminated at the F5 but we have the F5 send back to our CAS servers over SSL, too.

Ok, so let's go through a list of questions here to build a clearer current picture

 

 

Did you disable compression on LTM's HTTP profile?

 

Is the problem happening with all browsers(Firefox?) or just IE?

 

I assume this is OWA internal that is having an issue - but external OWA works fine?

 

External OWA that is handled by TMG - is TMG load-balancing directly to CAS servers bypassing F5?

 

 

 

Compression is still enabled on the HTTP Profile for the VS.

 

The problem occurs with all browsers.

 

Internal OWA is having an issue, external is not.

 

The TMG load balances directly to the CAS servers bypassing F5.

Can you please try to disable compression and see what happens?

Can't.

 

 

The OWA Persist iRule uses a COMPRESS command. I'm a complete novice when it comes to iRules. Is there a way to comment that out or should I just remove it and save? Note that we are running all Exchange services off of one VS.

 

 

when HTTP_REQUEST {

 

switch -glob [HTTP::path] {

 

"/rpc/rpcproxy.dll" {

 

switch -glob [HTTP::header "User-Agent"] {

 

"MSRPC" {

 

if { [HTTP::cookie exists "OutlookSession"] } {

 

persist uie [HTTP::header "OutlookSession"] 3600

 

}

 

else {

 

persist uie [HTTP::header "Authorization"] 3600

 

}

 

}

 

"*Microsoft Office*" {

 

persist uie [HTTP::header "Authorization"] 3600

 

}

 

default {

 

persist source_addr

 

}

 

HTTP::disable

 

COMPRESS::disable

 

}

 

}

 

default {

 

persist cookie

 

}

 

}

 

}

For testing, you can remove the iRule(s), HTTP profile, replace persistence with src ip based instead of cookie, and see if the issue still persists.

 

 

Thanks jdewing! I was having a similar issue and disabling the "Keep accept encoding" worked!

 

 

Hopefully it doesn't have any major effect on any other OWA operations.

 

 

Jason