The pending session on machine XXX for user DOMAIN\xxx has expired

[RTVNoord]

We have some strange behaviour on our Horizon environment. We have several desktoppools where people can connect to.
At the office everything works like a charm. Only from the outside (like home etc.) one pool is giving troubles; sometimes.

When connecting to this pool, the horizon desktop just keeps loading and loading.
In the Horizon admin interface you see no active connection on the machines in this specific pool.
In the events no error is shown either and only mentions that the pending session for that machine has ended.
The connection broker and uag's events, that are forwarded to our syslogserver are showing no errors 

When connecting to this desktop / pool the events is only showing this information

DOMAIN\bart
Warning
Agent
The pending session on machine W10-XXX for user DOMAIN\bart has expired

DOMAIN\bart
Info
Agent
The agent running on machine W10-XXX has accepted an allocated session for user DOMAIN\bart

DOMAIN\bart
Info
Connection Server
User DOMAIN\bart requested Pool win10-avid, allocated machine W10-XXX

DOMAIN\bart
Info
Connection Server
User DOMAIN\bart requested Pool win10-avid

The whole situation is not consistent. It works when on the Office network, and almost not when at home.
People can connect to other desktops / pools from home without any problems. It's just that 1 desktop / pool that is giving problems.

To make it even weirder: at home, sometimes people can connect from Windows desktops to this specific pool, but not from a MacBook. But not always. Some time later it also works on the MacBook. How weird. 

• We have 2 connection brokers (version 8.3.0-18294467) and 2 UAGs (version 21.06.2.0-18528989_OVF10)
• The Horizon clients is the latest for Windows and Mac
• The NVIDIA M10 GPU we are using runs at driver version 12.4 - 460.107
• The desktop with problems is a clone from other pools, is running the latest W10 21H2 fully patched and optimized.
• The Horizon desktops run on esxi 7.0.0

And yes, I've also deleted the pool and recreated it etc. etc. But the behaviour stays the same: unpredictable

What causes that users cannot connect to this desktop / pool from outside the network?

Any help would be great.

[Mickeybyte]

@RTVNoord 

How is the setup for external access?

• UAG -> LB -> CS? or UAG -> CS?
• Single of multiple UAGs? 

Be sure the check the session timeouts on all servers (UAG, LB, CS) to be sure they are aligned.

Also check LB settings to be as recommended by VMware (https://kb.vmware.com/s/article/56636)

[RTVNoord]

Thanks for your answer.

As mentioned we have 2 UAGs and 2 Connection Brokers
We also use a HA-Proxy as load balancer for the whole environment and all the external connections are following the same connection path.
We only use the Blast protocol

The timeout settings for the UAG, LB are the same (as far as I can see)
We also monitor the response thrue PRTG and that all is fine.

Just to mention: connecting and logging in from external, is not the problem.
After that the users do see all the desktops / pools that are available for them.

They also can log in to other desktops / pools, but only 1 desktop / pool (see attagement with the yellow marked one) is giving these connection errors. And it is not persistent, sometimes it works, sometimes not. But only with this specific pool. The Horizon client is just showing: loading desktop.

See attachement for the both images

Regards, Pascal

[Mickeybyte]

@RTVNoord 

Very strange. 

Try to list the differences between that 1 pool and the others and compare them. Could it be different network settings for that pool maybe? 

Don't know directly what this could be.

I would also suggest to create a LOG bundle and send it to VMware support for further troubleshooting.

[SurajRoy]

Is the Master image for the affected pool optimized?

Check the below settings on the impacted pool:

• Power setting under control panel should be set to high performance
• Turn of display and Put computer to sleep should be set to NEVER
• Refer to KB https://kb.vmware.com/s/article/1028975 and make sure the userint string is set correctly

[RTVNoord]

The only difference between the pools are the entitlements and network (different VLAN).
But if VLAN would be an issue, than it should never work. And thats not it. It worsk sometimes not and sometimes it does.

I've submitted a case @ VMware allready and awaiting their response. In the meantime I rely on the help off this community 🙂

[RTVNoord]

Yes, the master image is optimized same as all the other power settings etc.
I'm not familliar with that KB article and will check that!

Update: I've checked that registry string mentioned in that KB and that was ok. No modification needed.

[Tompous]

Hi @RTVNoord,

No updates about this problem?

I also have the same issue and reg is correctly configured

Thanks,

[RTVNoord]

@Tompous The problem is no longer with us. I can't even remember what the exact solution was 🤔

[Daniele_Palumbo]

Hi, we have the same environment with updated agent and software versions, would you happen to be able to remember how you solved the problem.
please we are going crazy