Forum Discussion
WVD - almost daily TermService crashes disconnecting all users
Hi team, A client of mine has a WVD environment compromising of a single Windows 10 Multisession host. This supports about 15 users, and runs really well for the majority of the time. The host VM...
Hi all,
Posting for anyone else who finds this thread. Unfortunately I jumped the gun with this one. While 21H1 has helped, we're averaging one Remote Desktop crash every 2 weeks now, which results in all the active sessions getting booted off into a Disconnected state.
It's better than it was, for sure, but there's still something fundamentally broken here.
Am continuing to work with Azure Tech Support on this one, but so far no real idea on the root cause or any deep troubleshooting has happened. I'll post back if I ever get to the bottom of it. Suffice to say, pretty disillusioned with both WVD and MSFT/Azure tech support.
Posting for anyone else who finds this thread. Unfortunately I jumped the gun with this one. While 21H1 has helped, we're averaging one Remote Desktop crash every 2 weeks now, which results in all the active sessions getting booted off into a Disconnected state.
It's better than it was, for sure, but there's still something fundamentally broken here.
Am continuing to work with Azure Tech Support on this one, but so far no real idea on the root cause or any deep troubleshooting has happened. I'll post back if I ever get to the bottom of it. Suffice to say, pretty disillusioned with both WVD and MSFT/Azure tech support.
Matt_Ignite I too had seen this behaviour earlier in the year but it also largely disappeared after we upgraded our session host VMs. However, over the past 3 days we have had a Term Service crash (svchost.exe_TermService) on 1 session host every day. We have 5 hosts in the pool and 1 host has been affected each day. This manifests itself in all user sessions on that host being disconnected. As in your scenario, the users can connect back in a few seconds later.
Did you ever get to a resolution on this with MS Support?
Hi tadhgclifford,
Yes, we did get a resolution in the end.
- I went back to Microsoft and the product team did more digging, along with reviewing all the dumps that we had captured (which was a lot, because it was crashing frequently.)
- Eventually they confirmed that it was due to a previously undiscovered bug (Bug 34382658) in the way the RDP stack handled multiple connections. Specifically "the known issue is influencing AVD connections because the inbox stack for RDP is sharing the same process with AVD SxS stack, and they are conflicting with each other. "
- In late September Microsoft released a patch to fix it - this one: September 30, 2021—KB5005611 (OS Builds 19041.1266, 19042.1266, and 19043.1266) Preview (microsoft.com)
- Since applying that patch in early October there's been no further crashes at all - it's been absolutely rock solid. Happy to report that it has resolved the issue completely.
Give that a shot - I'd be really keen to see if you've got that patch installed on any of the session hosts in your farm.
If it's related, we're on 21H1 with all the latest Windows Updates installed - completely bang up to date.
Good luck - let me know how you go.
Cheers,
Matt
Hi Matt_Ignite
Fair play to you for sticking with it and getting the resolution. Top marks for persistence!
As of today, we have applied all outstanding Windows updates to all our session host VMs (they only needed the most recent ones) so, hopefully, that will resolve it for us also.
Many thanks for your work on this and also your swift response.
Regards
Tadhg
Awesome work mate - fingers crossed that it fixes it for you. I'd have given anything to find an answer to it when it first started affecting us, so hopefully this will sort your machines out quickly.
Let me know how you get on with it!
Cheers,
Matt
