Web Interface Trusted Sites Issue

My company uses a Citrix Secure Gateway server with the Citrix Web Interface installed on the same server. If the address of the gateway is not in the end user's Trusted Sites list when using Internet Explorer, the user gets a message telling them how to add it.

However, since the Secure Gateway uses the address "localhost" to communicate with the locally installed Web Interface, the address the user is told to add to their trusted sites is http://localhost. This confuses some users, and if they follow the instructions it won't fix the problem. I needed a way to change the address to the actual URL of our Secure Gateway. Here's how to fix the problem:

1. Back up the file changeZoneHelp.inc (the default location is: C:\Inetpub\wwwroot\Citrix\XenApp\app_data\clientDetection\include), and open with Notepad. Add the following line after line 1 (replace with your URL):
<% viewModel.siteURL = "https://csg.company.com/"; %>
2. Restart the Citrix Secure Gateway service (this will close any open sessions).

That's it! The users should now see the correct URL.

Citrix "Protocol Driver" error when launching an application

One of our Citrix Xenapp 5 servers wouldn't allow applications to be launched, giving a "protocol driver" error. After following the instructions on Citrix's web site (http://support.citrix.com/article/CTX108698) and rebooting the server, the server started working again.

Here are the steps I took, in short:

1. Disabled session reliability for the server in the Citrix Delivery Services Console
   
2. Deleted the ICA-tcp connection in Terminal Services Configuration on the effected server
   
3. Recreated the ICA-tcp connection
4. Re-enabled session reliability
5. Rebooted the server

I didn't have to recreate the LHC like the Citrix article says.

Services.exe using large amount of memory

I noticed several of our Windows 2003 servers were performing poorly and found that the process "services.exe" was using between 200 and 400 MB of memory, compared to 6 to 10 MB on servers that were performing normally.

I fixed the problem by simply clearing all event logs. I have no idea why this worked, but it did and almost instantly brought memory usage back down to around 6 MB.