Hello guys!
We have pooled rds collection with virtual desktops settings, and a lot of problems with it, and we are looking for some solutions \ best practics from people who have it all works fine.
Configuration:
HA with two connection brokers in nlb cluster.
Four hyper v hosts in failover cluster with mapped san volue for vhdx
One file server for user profiles.
All of this working on 2016 windows, hyper-v hosts are physical, rest is virtual.
We use custom rdp client for pxe boot, so we have some debug information, that standart mstsc did not show.
First problem (and most common) is 0x0000405, ERRINFO_CB_SESSION_ONLINE_VM_WAKE. What about logs:
On connection broker
Microsoft-Windows-TerminalServices-SessionBroker/Admin
=== 10:10:47
RD Connection Broker failed to process the connection request for user domain\user.
Error: Element not found.
Microsoft-Windows-TerminalServices-SessionBroker/Operational
=== 10:10:47
RD Connection Broker received connection request for user domain\user.
Hints in the RDP file (TSV URL) = tsv://VMResource.1.VDI2
Initial Application = NULL
Call came from Redirector Server = RDSCB01.domain.local
Redirector is configured as Virtual machine redirector
On Host
Microsoft-Windows-TerminalServices-TSV-VmHostAgent/Admin
== 10:10:47
Remote Desktop Virtualization Host could not orchestrate the virtual machine with the GUID of A02F748B-51BF-44B0-B283-BA033CDF0A43.
Virtual machine GUID: A02F748B-51BF-44B0-B283-BA033CDF0A43
Hresult: 0x8000FFFF
We dont have auto save mode for desktops, (because it vm can not start in that state, and at start session user get error about "connection".) and have auto-revert vms option turn on.
Second problem it is that connection brokers is not disconnect user profile disk from vm after session get finish, so we have "feature" task, that every morning turn off all pooled vms and disconnect disks.
I have some conjectures, that when session limit expired, collection did not revert pooled vm to start position (even after forced stop-vm). So users when connect to not recovered vm, get their user disk mounted to this machine, but get error with start session and stuck in this condition for moment, when cb will not send them to another "good" vm.
Any solution for this? Another posh script for revert all machines is not good idea i think:)
About rdp errors:
https://msdn.microsoft.com/en-us/library/cc240544.aspx?f=255&MSPPError=-2147217396
We have pooled rds collection with virtual desktops settings, and a lot of problems with it, and we are looking for some solutions \ best practics from people who have it all works fine.
Configuration:
HA with two connection brokers in nlb cluster.
Four hyper v hosts in failover cluster with mapped san volue for vhdx
One file server for user profiles.
All of this working on 2016 windows, hyper-v hosts are physical, rest is virtual.
We use custom rdp client for pxe boot, so we have some debug information, that standart mstsc did not show.
First problem (and most common) is 0x0000405, ERRINFO_CB_SESSION_ONLINE_VM_WAKE. What about logs:
On connection broker
Microsoft-Windows-TerminalServices-SessionBroker/Admin
=== 10:10:47
RD Connection Broker failed to process the connection request for user domain\user.
Error: Element not found.
Microsoft-Windows-TerminalServices-SessionBroker/Operational
=== 10:10:47
RD Connection Broker received connection request for user domain\user.
Hints in the RDP file (TSV URL) = tsv://VMResource.1.VDI2
Initial Application = NULL
Call came from Redirector Server = RDSCB01.domain.local
Redirector is configured as Virtual machine redirector
On Host
Microsoft-Windows-TerminalServices-TSV-VmHostAgent/Admin
== 10:10:47
Remote Desktop Virtualization Host could not orchestrate the virtual machine with the GUID of A02F748B-51BF-44B0-B283-BA033CDF0A43.
Virtual machine GUID: A02F748B-51BF-44B0-B283-BA033CDF0A43
Hresult: 0x8000FFFF
We dont have auto save mode for desktops, (because it vm can not start in that state, and at start session user get error about "connection".) and have auto-revert vms option turn on.
Second problem it is that connection brokers is not disconnect user profile disk from vm after session get finish, so we have "feature" task, that every morning turn off all pooled vms and disconnect disks.
I have some conjectures, that when session limit expired, collection did not revert pooled vm to start position (even after forced stop-vm). So users when connect to not recovered vm, get their user disk mounted to this machine, but get error with start session and stuck in this condition for moment, when cb will not send them to another "good" vm.
Any solution for this? Another posh script for revert all machines is not good idea i think:)
About rdp errors:
https://msdn.microsoft.com/en-us/library/cc240544.aspx?f=255&MSPPError=-2147217396