RDSH App connectivity issue : No shared memory channel. Horizon client service may be down.

[davebaker87]

Hi all,

We have an RDS farm to allow some external staff to connect in, via a UAG to a few apps. For some reason the service stopped working, and now when we try to connect inbound from the internet, users see a black screen and the request times out. Connecting from the LAN it works fine.

The UAG seems fine (no authentication issue and desktop list populates in the horizon client). The error logs on the client machine (from appdata and \VDM folder) show a mix of errors as below (grouped).

We've tried using different protocol (in its working state PCoIP was the display protocol on the host), recovering the farm, restarting services, upgrading vmware tools etc to no avail.

Any suggestions on these? A google suggests that there's an AD issue looking up the computer account, and yes, I've read the vmware KB https://kb.vmware.com/s/article/2083784 - any other ideas!!??

2020-04-07T14:15:52.958+00:00 ERROR (453C) [WinCDK] PartnerAppMgr::ChannelAddRef : No shared memory channel. Horizon client service may be down.

2020-04-07T14:15:34.064+00:00 ERROR (453C) [WinCDK] USBDevices::FreeEnumData : CdkViewUsb failed to free enumeration. The reason is 'ViewUsbStatus_BadHandle'. The broker is 'smartext.uk.thalesgroup.com'. The desktop is 'CalculatorTest'.

2020-04-07T14:15:52.201+00:00 ERROR (453C) [libcdk] CdkClientInfo_GetMachineDistinguishedName: GetComputerObjectNameW failed with code: 0x00000547.

2020-04-07T14:15:52.201+00:00 ERROR (453C) [libcdk] CdkClientInfo_GetLoggedOnFQDN: GetUserNameEx failed with code: 0x00000534.

020-04-07T14:15:10.956+01:00 ERROR (4C54-9B38) <NodeManagerWatcher> [vmware-remotemks] Failed to initiate shared memory over RPC, server initiate failed.  1722 (The RPC server is unavailable.)

2020-04-07T14:15:10.956+01:00 WARN  (4C54-9B38) <NodeManagerWatcher> [vmware-remotemks] CORE::SharedMemChannel::Connect(): Channel (null) (0x0000000000000000): SharedMem Connect to SessionClient FAILED: RPC initiated sharedmem connect failed, reason=hostUnreachable

2020-04-07T14:15:11.456+01:00 ERROR (4C54-99A4) <39332> [vmware-remotemks] HccChannelPlugin::ChannelAddRef: No shared memory channel. Horizon client service may be down.

2020-04-07T14:15:11.456+01:00 ERROR (4C54-99A4) <39332> [vmware-remotemks] HccChannelPlugin::GetControllerMessageQueues: ChannelAddRef failed!

2020-04-07T14:15:11.956+01:00 ERROR (4C54-99A4) <39332> [vmware-remotemks] HccChannelPlugin::ChannelAddRef: No shared memory channel. Horizon client service may be down.

2020-04-07T14:15:11.956+01:00 ERROR (4C54-99A4) <39332> [vmware-remotemks] HccChannelPlugin::SendToAllControllers: ChannelAddRef failed!

2020-04-07T14:13:02.362+01:00 ERROR (7348-6F1C) <28444> [vmware-remotemks] HccChannelPlugin::ChannelAddRef: No shared memory channel. Horizon client service may be down.

2020-04-07T14:13:02.362+01:00 ERROR (7348-6F1C) <28444> [vmware-remotemks] HccChannelPlugin::GetControllerMessageQueues: ChannelAddRef failed!

2020-04-07T14:13:02.862+01:00 ERROR (7348-6F1C) <28444> [vmware-remotemks] HccChannelPlugin::ChannelAddRef: No shared memory channel. Horizon client service may be down.

[davebaker87 · Accepted solution]

To add more to this -

We have been able to successfully launch an app via RDSH to an external user but the reliability is intermittent. What seems to happen is that Pre-launch will commence shortly after authenticating with the connection server, then , if no pre-existing pre-launch session exists, it will try to create a new one. After exactly 1 minute, the 'new' session then expires. Attached image from our horizon logs:

UKDN.\S0088161test2	Info	04/09/2020, 2:39 PM	Broker	Get launch items timing profile for \S0088161test2 on XHZWIN0402, duration: 32
UKDN\s0088161test2	Info	04/09/2020, 2:34 PM	Broker	null application launch timing profile for UKDN\s0088161test2, duration: 124337
UKDN\s0088161test2	Audit failure	04/09/2020, 2:33 PM	Broker	Unable to launch from Pool a-uat-muapp-f1 for user UKDN\s0088161test2: There were no machines available that reported protocol [BLAST] as ready
UKDN\s0088161test2	Warning	04/09/2020, 2:33 PM	Agent	The pending session on machine AS12UATMuApp101 for user UKDN\s0088161test2 has expired
UKDN\s0088161test2	Info	04/09/2020, 2:32 PM	Broker	User UKDN\s0088161test2 requested Pool a-uat-muapp-f1, allocated machine AS12UATMuApp101
UKDN\s0088161test2	Info	04/09/2020, 2:32 PM	Broker	User UKDN\s0088161test2 requested Application a-uat-muapp01-goal_b1
UKDN\s0088161test2	Info	04/09/2020, 2:32 PM	Agent	The agent running on machine AS12UATMuApp101 has accepted an allocated session for user UKDN\s0088161test2
UKDN\s0088161test2	Warning	04/09/2020, 2:30 PM	Agent	The pending session on machine AS12UATMuApp101 for user UKDN\s0088161test2 has expired
UKDN\s0088161test2	Info	04/09/2020, 2:29 PM	Broker	User UKDN\s0088161test2 requested Pool a-uat-muapp-f1, allocated machine AS12UATMuApp101
UKDN\s0088161test2	Info	04/09/2020, 2:29 PM	Broker	User UKDN\s0088161test2 requested Application a-uat-muapp01-goal_b1
UKDN\s0088161test2	Info	04/09/2020, 2:29 PM	Agent	The agent running on machine AS12UATMuApp101 has accepted an allocated session for user UKDN\s0088161test2
UKDN\s0088161test2	Warning	04/09/2020, 2:28 PM	Agent	The pending session on machine AS12UATMuApp101 for user UKDN\s0088161test2 has expired
UKDN\s0088161test2	Info	04/09/2020, 2:27 PM	Broker	User UKDN\s0088161test2 requested Pool a-uat-muapp-f1, allocated machine AS12UATMuApp101
UKDN\s0088161test2	Info	04/09/2020, 2:27 PM	Broker	User UKDN\s0088161test2 requested an application session from Pool a-uat-muapp-f1, pre-launch true

I changed the pre-launch session timeout to 'Never' and tried to log in again and the audit failure at 2:33pm suggested no machines available, so would this suggest a licensing issue? Previosuly we've never seen this error because we didn't tweak pool settings.

Any guidance or troubleshooting ideas would be appreciated,

Dave

[epa80]

Hello. Did this happen to ever get resolved? We're experiencing something kind of similar in our environment, detailed in this thread below. Thanks in advance.

Re: Published App Fails on Initial Launch - VMware Technology Network VMTN