When there is more printing activity, printers mapped to a Citrix Universal Print Server goes offline. Alternatively, some applications might become unresponsive or fail when opening a print dialog box.
Recommended: Customers should upgrade to UPS 7.6.x, as the following is not relevant for current versions of UPS.
Citrix Technical Support has a code change that can be applied to expose a Registry Key allowing the permanent configuration of the number of XTE worker threads. The code change is not available for download at this time. Call Citrix Technical Support to acquire the change.
An immediate method to mitigate the issue is to edit the httpd.conf file on the print server. The caution to this method is that the changes made to the httpd.conf will be obliterated if either the server is rebooted or if the Citrix Universal Print Service is restarted.
Determine an average of how many printers per Citrix Universal Print server are mapped in a user’s ICA session.
How many users in all will be mapping printer to a given Citrix Universal Print Server.
Multiply #1 and #2, this gives you the number of threads needed. Use this value for the ThreadLimit value in the httpd.conf file.
For the ThreadsPerChild value, add 1 to the value used for ThreadLimit.
Certain print dialogs cause UPClient (the client side component of Citrix Universal Print Server) to make excessive connections to the print server. These connections eventually prevent the print server from taking on additional connections and any new printers appear offline.
To verify this is the case, go the XTE log file on the Citrix Universal Print Server. The log file is located at C:\Program Files(x86)\Citrix\XTE\logs\error.log.
In the log file, find one or both of these entries:The first entry is the warning indicating that the Citrix XTE Server is out of worker threads. The second warning with the (OS 64) error is a report of a WinSock error caused by a failed attempt by the client to connect to the service. There might be many of these entries - each entry corresponding to a failed connection attempt.
It is important to understand that both warnings must be present for the symptoms to synchronize with the cause.