Problem Definition:

Multiple Partners (i.e. more than 4) concurrently report end users are unable to access their desktops.  End users experience a black screen that lingers when logging on.  Or end users complain their logon hangs or lingers at the logon screen.  This generally presents early in the morning.  Restarting the affected customers’ desktop hosts does not resolve the problem.  Users do not have an issue authenticating against the Citrix Web Interface or NetScaler.

Cause:

A Windows or VMware Tools network driver update may cause customer domain controllers to lose connectivity with their primary network adapter.

Resolution:

WARNING: THIS RESOLUTION ASSUMES AN EXCESSIVE NUMBER (i.e. MORE THAN 4) CUSTOMER DOMAINS REPORT SERVICE INTERRUPTIONS AND IT IS IMPOSSIBLE TO KEEP UP WITH THE OUTAGE REQUESTS.  THIS RESOLUTION WILL RESET ALL CLOUDCONNECT DOMAIN CONTROLLERS.  IF ONLY A SMALL NUMBER OF PARTNER DOMAINS ARE AFFECTED, APPLY THIS RESOLUTION ONLY ON THE AFFECTED CUSTOMERS’ DOMAIN CONTROLLERS AND DESKTOP HOSTS.  ONLY USE THE FOLLOWING PROCEDURE IN AN EMERGENCY WHEN THE ESCALATION ENGINEER IS NOT AVAILABLE.


Restart all domain controllers, wait 3 minutes, then restart the affected desktop hosts.


From the vSphere Client, highlight/select the “PRODUCTION VMS” Folder.  Click the Virtual Machines Tab.  You can quickly locate all domain controllers by searching for all virtual machines for a virtual Machine Name matching the following strings, note that each string will require its own search and virtual machine reset operation “-D1” “-D2” “-D3”…“-D9” etc. (do not search with quotations).  Sort the VMs by Power State.Left click the mouse and hold, drag the mouse across all VMs listed.  Once all VMs are highlighted, right click anywhere on the highlighted VMs, hover over power, and in the sub menu, choose reset.  All CloudConnect Domain Controllers in the “-D1” group will now reboot.  Now search for “-D2” and repeat this procedure until no more domain controllers are visible in your search.


Note, if not all customers are affected, only located and reset the Domain Controllers of the affected customers reporting service outages.


Next restart the affected customer desktop hosts as needed.  If it is clear that all customer desktop hosts are affected (this would be very rare), you can quickly locate all desktop hosts by searching for all virtual machines with a virtual machine name matching the string “-H1” “-H2” “-H3”…“-H9” etc.


Additional Information:

Use this Article’s Resolution with great caution and as a last resort only when all indications match, and the Escalation Engineer is inaccessible.

Applies To:

Citrix XenApp 6.5

VMware vSphere 5.5