20 to 30 second black screen on reconnecting with PCoIP and Blast

[billmaxpowers]

   We are having an issue with a 20 to 30 second black screen on reconnecting a active sessions, in both Win 7 and Win 10 desktops. It happens on both PCoIP and Blast. We tested RDP with no issue, it connects right away. This issue has been going on awhile but most users are login off and locking back in. There is no delay with that, it connects right a away and you see the login process.

[RandyBrown]

I've been working with VMWare support for weeks on this one.  They finally have been able to reproduce in their lab.  However, I received a call yesterday from them to inform me that they consider this "expected behavior" with Windows 10 on Horizon View. 

We disagree that this should be considered expected behavior and are in the process of setting up a meeting with our VMWare rep to discuss further.  One of the main reasons that we were able to deploy VDI was because of the improved performance at login.  This takes us back to the days where users had to wait 10-20 seconds to login to their computer ... UNACCEPTABLE.

RyanChristopher​ ... to determine if it's memory usage, just open up Internet Explorer with 10 tabs, open some Word and Excel documents, maybe a couple PowerPoint documents ... until you have ~60% memory usage.  Then disconnect and reconnect.  See how long that takes.  Then increase your memory usage to ~80% or higher and try again.  Notice how much longer it takes to reconnect.  It is a direct correlation to memory usage.

[sjesse]

I'd escalate your case, the last time I've heard of this it was an appvolumes bug, I'm in the process of testing a new upgraded enviornment so I'll keep an eye out for this.

[RyanChristopher]

@RandyBrown For testing I actually opened up as many apps and browser tabs (both IE/Edge) as possible. Again with 8GB RAM it was hard to get over 50% usage here so perhaps we aren't running into this issue. I can back our image back down to 4GB and test again. We had bumped up RAM for our pilot to rule out some other issues we were having with performance degradation when power users would have all their apps/browsers open.

If we test reconnect times with nothing running and just at the desktop we'll see around ~10 seconds, up from our Windows 7 pool with ~5 second reconnect times. Again, this using PCoIP/BLAST.

Wanted to also mention we saw improvements when removing Imprivata from our image (or disabling processes/services temporarily) and reconnecting. Not sure it this is because of Imprivata/VMware/Windows10 but we have a case open with Imprivata now to investigate further. Will post any updates here.

[JohnTwilley]

This may sound weird, but I've seen where removing the "Lock" option from Ctr+Alt+Del will cause a 20 - 30 second delay on logon or reconnects in Windows 10 18XX versions.

No, I don't have a source. But I saw and fixed it myself using this method & GPO as well, cause "why not".

Try Disabling the option to remove it in Group Policy, or via local PC registry keys (at your own risk,,,LOL)

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Winlogon]
"DisableLockWorkstation"=dword:00000001

[JohnTwilley]

Forgot to add a shout-out to this post that got it going: Re: CTRL-ALT-DELETE screen on reconnect

[iSeb2]

Got the same issue, for the moment I use the "DelayedDesktopSwitchTimeout = 15" workaround, which means that we see "Preparing Windows" instead of the black screen (total logon of about 30 seconds), but still, this behavior is not really acceptable.

[ToTanne]

Hy together.

Same problem here. 20 seconds login time with WIN10. We haven`t had that before on WIN7

About 2000 VDI Desktops running

- WIN10 - 1809

- Horizon View 7.9

- Imprivata 6.3

- No GPU

- No AppVolumes

- PCoIP

I hope there will be a solution in future

[VMwareAdmin123]

I have a case open with Imprivata about this.  Try downgrading your Imprivata agent to 6.0 HF10.  The reconnect times should get you inline with Windows 7.  I figured this out by trial and error, any non 6.3 version of the Imprivata agent doesn't seem to exhibit the reconnect issue.  This is obviously a stop gap solution until this gets resolved in the future.  All my Imprivata SSO profiles still worked after downgrading to 6.0 HF10.

[Techstarts]

@mchadwick19   

It is old thread but hope you can still help me.

My image is not joined to domain.

Now I created a manual pool.

I connect using PCoIP and what exactly you mean by resizing.

we have two monitors. Shall I resize to full screen?

Thanks in advance

[RK06]

Did anyone resolve this issue? We are experiencing the exact same problem. Seems to be a 12-20 seconds delay on reconnect. RDP is instant.

- WIN10 - 1809

- Horizon View 7.11

- Imprivata OneSign 7.0

- No GPU

- Blast

The issue also occurred on 7.7.

[maharajan_be]

Anyone resolved this issue. We are also experiencing the same problem "black screen on reconnecting existing vdi session" Win 10 1909 Horizon view 7.11 UEM 9.10 No GPU

[Melandrach]

I know this is an old thread, but we have the same issue with Windows 10. We never had this issue previously on Windows 7, reconnects were always very fast to the desktop display (3-5 seconds at most)

- WIN10 - 1809

- Horizon View 7.5.4

- Tested with and without Imprivata agent

- No GPU

- Tested both Blast and PCOIP

We have opened a case with VMware and they have initially told us this is a known issue with Windows 10 and have kind of blamed Microsoft for it, even though it only happens on VMs. We also now have a case opened with Microsoft. I'll update this thread if we make any progress.

[RK06]

We had resolved the issue.

The issue was related to Imprivata OneSign. The OneAgent had a bug causing the black screen delay, we were provided with a custom OneSign agent that resolved the issue. We were also informed that the fix was included in the following mainstream release. 7.2 SP1.

We have not had a change to upgrade to 7.2 SP1. So I am unable to confirm if the issues no longer occurs under 7.2 SP1.

[JohnTwilley]

Which Imprivata Client had the bug?  7.2?

[RK06]

We had identified the issues during our initial Horizon pilot. We would receive the 10-15 second delay after we added the OneSign agent to our master image.

We had tested various version of both the Imprivata Agent / Appliance. We worked with VMWare/Imprivata TAC and identified the issue was with the OneSign agent. We finally settled on the following and was provided a custom OneSign Agent. That was installed on our Master Image that resolved the issue.

I was recently informed that the fix was applied to 7.2 SP1. I am not sure if it was included in any of the HF patched for 6.3, 7.0, 7.1

We have not upgraded yet to 7.2 SP1. We have this on our road-map for October.

Production Versions:

vCenter 6.7 U3I  - 16616482

vSphere ESX 6.7U3-16316930

Horizon 7.11

App Volume 4 - 2006

DEM 9.11

OneSign Appliance 7.0 HF2

OneSign Embedded 7.0 HF2

OneSign Agent 7.0 HF2

OneSign Agent VM Master Image 6.3.199.704 (I was informed this was a custom release)

[TurkeyVDIcall]

I'm running a similar environment, and am battling the same issue. Do you have any information you could pass on for the custom Imprivata agent and how one goes about getting it exactly?

[Melandrach]

Did you have this issue on VMs even without the Imprivata agent?

[RK06]

Send you a DM with the ticket number you can reference.

[RK06]

Without the OneSign agent, we did not experiance the black screen under Blast, RDP and PCOIP.

With the OneSign Agent - Black screen under Blast and PCoIP, RDP was fine.

[catalinuxmm]

Hi Melandrach, were you able to figure it out? I am experiencing the same issue here using HV 7.5.3 with win10 1809, no OneSign agent. Thanks