Hey guys,
We have started rolling out app volumes to some test users, and have gotten back a laundry list of issues. Any help or recommendations are appreciated. We are using writable volumes (UIA+Profile) and a few appstacks on instant clones. We're on relatively current versions of everything (either current or one version behind). We are not using UEM, though I've seen several people saying that we should. I don't entirely understand why we need it, since we don't push a lot of settings out to our users. We have a very small IT team so the fewer tools we have to manage the better, which is why we went without UEM. If UEM is actually required for a stable environment that would definitely make us look at using it again though.
So here's a rundown of the issues we're running into.
1. Users are unable to delete folders using the recycle bin. Bypassing the recycle bin works, but isn't really a good solution. From what I've found this seems to be an issue with the recycle bin when mountpoints are being used. Has anyone found a way around this?
2. The default web browser reverts back to the default set on the golden image at each login. The default on our golden image is chrome. If a user changes it to firefox it is lost the next time they log in. I've tried adding some includes for registry keys I found on the snapvol.cfg but no luck.
3. Default printer keeps getting lost. A user changes the default printer and the next time they login it has reverted back to the default on the golden image.
4. Lots of application settings get lost. Things like view settings, that stick on persistent desktops are lost on each login.
5. Addons for office are getting disabled at each login. Users have to reenable them every day.
6. Evernote has loads of issues. The desktop and start menu shortcuts created when it is captured in an appstack are all screwed up, so users have to launch the executable from program files. For some users it just won't work at all and throws an error about a "missing entry point".
7. We've had a couple of instances where a users writable vmdk was deleted when their instant clone reset. It seems like there should be some safegaurds to prevent this but I've seen it happen twice and we had to recover the vmdk from a snapshot on our SAN.
8. We periodically have an issue where writables and appstacks will stop attaching. It typically starts with one or two users having failures, but they're able to logoff and back in and everything attaches. It will then progressively get worse until they stop attaching for everyone. After a few hours of this everything starts working again. No errors in the app volumes console. When this is happening I am able to see that the reason it's failing is because of the known issue with the "Manager1" registry key getting deleted. The strange thing is that we have a gpo in place that support had us set up that is supposed to add the key back if it's missing. When these failures start to happen though it doesn't have any effect, even though all other gpo's seem to be working, and running gpresult shows that it is being applied. This isn't from any type of boot or login storm. We've only got about 20 users testing these desktops out at the moment.
I don't suspect issues in our view or vsphere environments in general, primarily because we've been running horizon view for several years, just using full clones. Until we started using app volumes everything was very stable.