I'm hoping there are some good ideas out there on how we might find the source and solution to our problem. This issue has us stopped in our tracks. We use full clones and we are in the process of doing in-place upgrading of Windows 10 from 1803 to 1809. We have deployed fresh installs of 1809 without any problems. Our initial testing of the in place upgraded VM's looked good but now that we have heavy usage on them, we have found a problem. VMware support has indicated that they only suggest fresh installs and to not try in-place upgrading but we really need this ability.
Setup:
- Horizon 7.8
- ESXi 6.5
- Full clones with persona management and folder redirect
The problem presents itself as the following...(this problem could happen once a week or as bad as multiple times a day)
- If the user is actively on the VM, they notice what appears to be a slowing down of their VM. Programs like Excel may crash, Chrome eventually indicates that it cannot access the Internet and they may get to a point where they cannot open programs or get windows to respond correctly.
Things we have noticed...
1. This is random. We have only been able to reproduce it ourselves once. We left Chrome going with multiple website (CNN, leakspin, Gmail) and a couple days later Windows was misbehaving.
2. vSphere still shows that VMTools and the View Agent as "running".
3. We are able to ping the computers IP address and usually HTML or Horizon client connections still open the VM.
4. Windows does not respond to much. For example, we can send Cntrl-Alt-Del and get the windows blue screen to come up but we cannot get Task Manager to open.
5. A remote systeminfo command shows available memory and CPU as fine.
6. When attempting to open a couple programs yesterday, it took hours before the VM brought the program up.
We took a snapshot of the Test VM that started doing this, encase it is possible to recover it to a broken state. If we shutdown and startup the VM, it will behave correctly for an unpredictable number of days.
Thoughts on how we might find the needle in a haystack on this one?
Are the problem desktops running on a VMFS5 or VMFS6 datastore?
What AV do you have on the machine? Is it compatible with 1809? off topic, if you are using full clones why do you have persona and folder redirect?
We use VMFS6
I assume by AV you mean VMware Agent Version? We are using Horizon 7.8 so our agents are 7.8 with our 1809 VM's. They were fresh installed with our new builds and we upgraded from 7.5.1 to 7.8 (un-install / re-install) on the machines we did in-place upgrades on.
We use persona management/folder redirect to benefit from the ability to use folder redirect. We point our users My Documents, pictures, etc. onto a file server. This makes for giving a user a new VM a little easier than having them start completely over because they don't loose access to important saved content.
On Friday we received the first report of someone having the exact same problem with a fresh Win 10 1809 build so we are taking a step back and setting up several test machines in an attempt to try to reproduce the issue and to see which combinations have the problem. Our test machines are starting today and account for the following scenarios.
- Fresh install Win 10 1809 VM with 7.8 agent
- Fresh install Win 10 1809 VM with 7.5 agent
- Win 10 1803 VM with 7.5 agent (what our previous VM's have been using)
Our test will the have the following items running on each VM.
- Empty Excel document open
- Chrome open with the following tabs (Gmail, Leakspin video loop, CNN, Fox News, ESPN)
Pretty sure he meant antivirus, if you have an antivirus try disabling it before the in place upgrade.
Ah, yes. We use Sophos. Since we are starting to hit the problem with our fresh VM's I'm not sure it is related or not. Once we get through our testing and verify if there is any pattern, we may want to remove AV from the machines to rule it out as well.
I haven't come across it myself, but have seen mention of performance issues related to running 1809 on VMFS6 datastores. Supposed to be fixed in 6.7 U2 and 6.5 U3, but haven't seen anyone posting about it being fixed or not.
This problem has been confirmed to happen on fresh built 1809 VM's as well.
We have made some progress on narrowing the possible cause to the issues we are experiencing. We believe the problem revolves around or relates to the VMware Persona Management service. Once we hit a VM with the issue, we have been able to stop the process "VMWVvpsvc" and it then it begins responding correctly again. The VM will go from a state of not responding (cannot open or close programs) to immediately getting caught up and responding well again after that process is stopped. This has been confirmed on each VM we have tried it on since this discovery.
Anybody else have ideas on how to address the VMware Horizon Persona Management causing this issue on Windows 1809 machines? This could be a bug.
Note: We use Windows Persona Management and Folder Redirect so we assume this service should be installed?
We have a couple test 1809 VMs that are using View Agent 7.5.1 and they have yet to have an issue but it could be just a matter of time since the problem is very sporadic.
Use VMware User Environment Manager instead of Persona Management. VMware gave up on that solution for a reason.
familylifetech,
As cbaptiste said, you're using Persona for folder redirection, but it would be much better to configure UEM for folder redirection instead. There's no need to have both solutions installed on the same system unless you've just started migratng from Persona to UEM.
Lars
I am just starting to read up on UEM as I am not familiar with it. I see a post from 2017 that talks about how to migrate people.
Would UEM cover our use case?
- Full clone VM's
- Users are admins so they can install any program so we do not use thin apps or app vols, etc.
- We want the ability for folder redirect of my documents, pictures, etc. The main use case is so that we can give a user a new VM (when they break there current one) and there personal files follow them to the new VM.
Think of FSLogix. It just works out of the box. I think you (your company) are also entitled for it as it is free since MS has acquired it.
FSLogix has disrupted the Roaming profile. I see you have perfect use case for it.
I have no relations with FSLogix. Im speaking from personal experience. Moderators can check my earlier threads.
familylifetech,
Basicly what you're asking for here is folder redirection. UEM is indeed capable of doing that and much more.
Lars
Yes you can manage user personalization on your instant clones, linked clones, full clones and even physical desktops with User Environment Manager.