Horizon 7: UEM and The Case of the Slow Logoffs
Once upon a time, there was a client … they lived in an office complex on the edge of the forest and had a new Horizon View 7 environment complete with the latest User Environment Manager (UEM), App Volumes and Horizon for vROPS monitoring.
While we were on site working with the client we built instant clone pools and used UEM 9.4 to profile applications for a variety of use cases, enabling the client to realize their business and IT drivers:
- Increase the efficiency and effectiveness of technologies used by the company in support of end user devices
- Support end user customer satisfaction for use of current technologies in performing tasks within the IT infrastructure; PDAs smartphones, tablets, etc.
- Reduce the amount of help desk support required to maintain end-user devices thereby reducing operating costs
- Reduce the high cost of end-user device deployment and upgrades
The infrastructure was designed, built and instant clone desktops were deployed. While deploying and beginning testing on these instant clones it was observed that, on occasion, the logoff process would hang preventing the user from logging into another session. This hanging condition would last 10 minutes when, eventually, the Windows logoff process would time out and force the session to close.
This proved to be no easy problem to solve. The first challenge was with instant clones. On logoff, they are deleted from the vSphere environment and a new instant clone is forked, creating a new instant clone desktop available for users. We needed to gather logs from the hanging instant clone desktop. In order to do this, we needed to prevent the instant clone from deleting on logoff. See this vJoeG blog post for the details.
Once we had the logs from the hanging instant clone desktop we shared these with the VMware Global Support Services (GSS) team. After a number of trials and errors, we found that the spooler.exe service was hanging at logoff due to User Environment Manager trying to un-map the network printers present in the session. This would not be a big deal except the print queue names contained spaces which could cause issues as this Microsoft article points out. One of the client engineers had the print queues updated to remove spaces and the hanging condition went away, there was much rejoicing. 
This is just one of a number of times unforeseen powers can be at play affecting your own or your customer’s environment. Keeping a cool head and reaching out to as many resources as you can enable you to quickly and effectively get to the bottom of your problem.
Prior to contacting VMware Global Support Services (GSS) we also leveraged the VMware communities, another great source for collaboration and support.
