VM Connection very often gets disconnected

Copper Contributor

Hello! 

I'm seeing a lot of disconnects in my hostpools. 

Just in the last 24 hours, a user had the following

The connection from the client to the Windows Virtual Desktop service was terminated unexpectedly. (24 times)

The network connection between the Windows Virtual Desktop client and the service was unexpectedly interrupted. (50 times)

 

So far the Azure infrastructure is good. No issues with the session hosts. This issue is affecting 20 users out of 700. 

Is there any way I can investigate deeper to get more knowledge on what is going on? The LogAnalytics is only giving partial information

62 Replies

@nicolasriderelli Intersted to find out whether this ressolved the issues. We are experiencing simialar problems with our WVD deployment

Either as a machine or user based install, the MSRDC doesn't exist for us in HKLM. It exists in HKCU but not in HKLM. Would this not also work in HKCU?

Thanks

Hey all I'm seeing the same issue with a small % of my users. Like previous comments 95% work just fine but ~5% have irregular issues. This is a disconnection so my users can reconnect and get their same session back and it only happens to a few on the same session host.

 

Here are the Errors and Sources.

SourceErrorMessage sample
RDGatewayConnectionFailedClientDisconnect (-2147467259)The network connection between the Azure Virtual Desktop client and the service was unexpectedly interrupted.

 

Things Checked:

  • User is connecting to closest region (Central US)
  • RD client is up to date (windows)
  • Heartbeat registry entries put in image to prolong connections (unsuccessful)

Testing:

  • Web access for a week for impacted users to see if that will calm down the disconnections
  • HLM\Software\Microsoft\MRSDC\Policies REG_DWORD AutomaticUpdates (2) on impacted host pool

Testing will take a couple of weeks because I do not want to make two changes at the same time. If it works and I make both changes we won't know what fixed it. I am thinking this is an issue with user at-home network or the RD client. Do we have any old versions of the RD client that seem to be working better?

Hi all, Been having this issue since deploying for local clients, we can go through days with no issues, and then all of a sudden we get a lot of errors. We have a mix of Windows-based and Linux based client devices ,also some using the web client and all devices affected. When it occurs the users can be across different hosts on different devices however not all users on hosts are affected. We have a ticket open with Microsoft but this has been open and ongoing since Feb with them. Its a pain to the client but nice to see other users are experiencing the same issues. 

Hi.
No. So far no. Im trying with shortpath in our environment. That lowers the amount of disconnections per user but some of them are being disconnected randomly.

Thought I would bump this thread as it has been a continual problem for us. Users will just get randomly disconnected from the session. It just closes without warning, the user can then reconnect back to the session, we hoped that the deployment of shortpath may help but it doesn't look like its made a difference. 

 

Deployment RDStack ShortpathTransportNetworkDrop (68)

 

We have gone to our network team who have continually said bandwidth is fine and no issues on the firewalls, all AVD URL's are accessible. 

They came from Citrix which seemed to be more reliable and the have questioned about going back which is a shame.

 

 

@KevHal 

 

Are you using Windows 10 or 11 on AVD? I had so many people face disconnects using Windows 11, but when I rebuilt the VMs using Windows 10 the problems stopped happening.

Hi,
No these are all Windows 10 Session Hosts. The session response is absolutely fine, no latency, it just cuts out for no reason. So disheartening.

@CelDB windows 10 hosts , not all sessions really random 

Did you ever find the cause or get a solution for these disconnects?

@ATWVD No never got to bottom of it 

This error message suggests that there was an unexpected interruption in the connection between the client device and the Windows Virtual Desktop service. This could be caused by a number of factors, such as network connectivity issues, problems with the client device, or issues with the Windows Virtual Desktop service itself. To resolve this issue, it is important to first identify the cause of the problem, which may involve checking network connectivity, troubleshooting the client device, or checking the status of the Windows Virtual Desktop service. Depending on the cause, potential solutions could include restarting the client device or the network, or reaching out to the Windows Virtual Desktop service for further assistance.
Also been still having the same problems, ticket with Microsoft has been going on for +- 4 months now. As of now we still have no resolution and the clients that are bothered by it are close to stepping away from AVD because we can't get to the root cause of this issue.
Hi Kugan
I have spend almost 9 months troubleshooting with Microsoft support. The major explanation was network issues with the end user.
The thing is, we started troubleshooting and then the user had no more disconnections and a new user came with the problem. It was very complex to troubleshoot since is random. One day you can have the issue and the other you are good.
Even with shortpath for manage networks the problem persist. We enabled Microsoft VPN with a high bandwidth, and the user see a little stability but still disconnecting. This is interesting, because when we eliminate the gateway, the problem is less frequent.

Thanks!

@nicolasriderelli Hi, the client has been using ping plotter across two ISP providers one fiber connection and one VDSL. The office mainly use the fiber with either shortpath via vpn for windows devices and over  internet for linux based clients. On the fiber they have seen significant packet loss accessing microsoft. The ISP however says its not their problem and is with Azure services. Testing this against the VDSL line they say they get a lot less packet loss. This is done at all times during the day. We spent about the same time as everyone else with this issue about 9 months with MS collecting logs but never got to the bottom. Seems both parties blame each other. We did improve things apparently when moving some file shares from a virtual server to azure file share , however not sure how much the users are reporting disconnects. I suspect the issue is with ISP's rather than MS and interested to hear other peoples opinions

Have you ever managed to get to the bottom of this issue?

HI @KevHal, This is the solution, for all facing disconnections, the new RDP client comes with a by design addition of rdp shortpath, you need to disable it,

I suggest to this 

You can easily disable the feature by disabling the following GPO: Enable RDP Shortpath for managed networks
It is under Computer Configuration > Administrative Templates > Windows Components > Remote Desktop Services > Remote Desktop Session Host > Azure Virtual Desktop.
More information, here: Configure RDP Shortpath - Azure Virtual Desktop | Microsoft Learn
Computer Configuration > Administrative Templates > Windows Components > Remote Desktop Services > Remote Desktop Session Host > Connections.
Select RDP transport protocols. Set it to Enabled, then for Select Transport Type, select Use only TCP.
Gpupdate /force after thatn and reboot the server just in case.
Hi @tejas_memane, thank you for this finding! It seems that in the AVD insight I found Shortpath network issue log which is related to the disconnection error.
Did that end up working for you? Weve been hearing a lot of chirping about this and I too am seeing the error (68). I have put in the GPO with EndPoint to a few users to test.
I put that change with the Admin Template in EndPoint but it still doesn't show reflected on the computer when view through Local Group Policy on the machine. Is that normal?