We are experiencing a really odd issue in VMware Horizon 7.4 and Windows 10.
Our Horizon infrastructure is on VMware Horizon 7.4
Client agents are on 7.6 (due to issues with Persona not working correctly with 7.4)
Client OS: Windows 10 1803
We are not running app volumes
Order of events:
1. User logs on via PCOIP and connects to workstation fine.
2. User disconnects and reconnects at another time and they are faced with a black screen (intermittent)
3. Session is completed unusable and unable to logout.
Checking vCenter, the CPU and the VM looks to be pinned however after logging on to the VM using the local admin account (took 90 mins to login, the CPU is not pinned.
I've raised a case with VMware who stated that we should put in a task to logout the workstation completely at the end of the day which is not really an acceptable answer. Thankfully, we are in pilot at the moment but this issue is starting to annoy me.
Does it matter too much if I run 7.6 agent on the workstation and 7.4 on the Horizon server components?
I've removed the agent and tools and installed them in the correct order. Now I've just ran out of ideas.
We had an issue earlier this month with the Win 10 App Readiness service (we're also on 1803) going crazy for an unknown reason. There are lots of articles out there with similar complaints and the solution for us was to disable the service. So far, we have not seen any negative impacts of doing so. It sounds like you're not able to get into a VM to see what app/service is spiking the CPU but wanted to mention it as a possible starting point.
I'll check that service but I've got a feeling it's disabled.
I forgot to add that I've applied the login.vsi template using the VMware optimisation tool.
We were running Horizon Agent 7.7 with Connection Server 7.4 for a longtime to address a bug.
Which version of VMware Tools are you running? I would recommend a uninstall of Horizon Agent/VMware Tools/any other VMware components. Then do a fresh install of the latest supported tools in the correct order.
I'll give that a try! We were leaning towards updating the agent again to see if it resolved.
We are running 10.1.15.
I had that issue as well on my 1809 image running the Horizon 7.6 agent...it ended up being a Group Policy that hide the Lock option from the CTL+ALT+DEL screen.
Seriously taxed my Google-Foo skills. And why does it cause a long black screen pause...? No idea. But it fixed it.
Just do a CTL+ALT+Insert on your VM and see if LOCK is missing!
If it's that, we can all laugh together...
So... just an update to this issue.
I installed Horizon 7.7 agent and recomposed the pools but unfortunately the issue still occurs. I can confirm there is no GPO to remove the Lock option on the lock screen,
We did manage to get the workstations responding again by restarting the VMwvvpsvc.exe service which I think it the Persona service. It's unbelievably frustrating.
Any update to this issue? We are experiencing this as well.
Horizon Agent 7.11
Windows 10 LTSC
We are experiencing this issue with Horizon Client and ThinOS thin clients and it appears to be on a non-graceful disconnect. Also occurs on external clients coming in through UAG and internal clients pointed to Connection server direct.
* Problem is VERY intermittent
* Horizon 7.9 with VM's on Horizon agent 7.11using Blast
* Horizon Template VM and pool video settings per best practices and Tools/agent/UEM installed in correct order
* The VM is accessible through UNC and management tools
* Windows 10 1809 LTSC
Black screen cause can be many things, and the general approach to troubleshooting this is well described here:
However, if you see the CPU spiked and cannot get into the machine even through the console, then I recommend using a professional EUC performance monitoring tool like Liquidware Stratusphere UX or ControlUp.
Using an EUC performance monitoring tool, you can look at historical data on a specific date/time on what CPU process was spiking, and what other activities were going on at the time -- like GPU spikes, network or disk IOPS. Correlating those activities with the CPU spike can give you the answer.
For example, a common cause of a CPU spike is antivirus software not having proper exclusions for Horizon processes/files and thus scanning the processes/files. You will get a CPU spike from the Horizon services and IOPS activity from the antivirus.
Check that you're using the SVGA driver included with your Agent. We just resolved a black screen issue by removing the SVGA driver that Tools was laying down. https://kb.vmware.com/s/article/2078739