Windows 10 Pool Disconnects and Guess OS Crash

[cdubz]

Have a windows 10 linked-clone floating desktop pool that is experiencing frequent disconnects on dell wyse zero clients (p25 and p45). When the disconnect happens the guess OS crashes, resets, and registers the event in vsphere. VM is then placed in an already used state in View.

This is the only error in the PCOIP log I can find.

SERVER :map_agent_to_tera: DISCONNECT_FOR_RECONNECT -> TERA_DISCONNECT_CAUSE_HOST_BROKER_RECONNECT

2016-06-07T07:43:34.220-04:00> LVL:0 RC:   0       MGMT_SESS :Tearing down the session

2016-06-07T07:43:34.220-04:00> LVL:1 RC:   0           VGMAC :Stat frms: R=000000/000000/000000  T=000000/000000/000000 (A/I/O) Loss=0.00%/0.00% (R/T)

2016-06-07T07:43:34.220-04:00> LVL:2 RC:   0          COMMON :TERA_PCOIP: SESSION_EVENT=TERA_MGMT_SYS_SESS_EVENT_RESET, disconnect cause (0x105)

2016-06-07T07:43:34.220-04:00> LVL:2 RC:   0          SERVER :server main: cb_notify_session_status called (mask 0x10) with tera_disconnect_cause (0x105)

Connection Servers are sitting behind a F5 virtual NLB.

Anyone else run into issues with Win10 crashing in a VDI environment.

[cdubz]

An update for anyone else trying out windows 10 in VDI.  We are still having the issue and have pretty much run down the cause but have not found a fix action yet.  What is happening is the guest OS (Windows 10) is full on BSOD crashing at random times and intervals when users are connected to them.  When it happened users would just get a disconnect message.  On the back end the guest VM would reboot itself and be put in an Already used state.  We were not ever getting an actual dump file of the crash because it looks like the vmware optimization tool disables all that.  Do after going back to the master image and enabling it again and then do a recompose we were finally able to capture a memory dump of the BSOD happening.

Running the dump through debug and the OS is crashing with the Persona Management driver trying to write to unallocated address space. We even tried upgrading the whole view environment from view 6.2.2 to Horizon 7.0.1 and the problem persists even on the newer agents and driver sets. So has anyone successfully gotten windows 10 to work consistently and stable with View Persona Management?

[h3nkY]

There is a bug fix on 7.0.1 similar to this issue but not included in 6.2.2.

The plan will be in 6.2.3.

[h3nkY]

I'm pretty sure the blue screen crash issue has been fixed in 7.0.0 (7.0.1 as well).

If you see the issue with 7.0.1, please raise support ticket and mention this thread.

Thanks.

[cdubz]

Thanks for the feedback.  We do indeed have a support ticket open with vmware at this time.  And the issue is definitely still happening in the 7.0.1 environment which is what we are running now.

[h3nkY]

Please prepare master image VM where sometimes technical support will ask you.

[h3nkY]

Just to ensure that you are using Win10 enterprise not professional version.

[cdubz]

roger that we are indeed using windows 10 enterprise on the 1511 MS release.  and we have actually tried two other pools with just the basic os/patches, vmware tools, and view agent w/persona managment and still have the issue happen.

[eccl1213]

Have you had any news from VMWare on this?

We just deployed a Win10 desktop pool on View 7.0 and we have the same issue.

The VDI desktop blue screens and reboots.   We changed the VM to create a dump file and dumpcheck shows that the issue is most likely Persona (VMWVvpfsd.sys)

We did start a support ticket but haven't heard anything back yet.

[cdubz]

No 100% fix yet. They mentioned it might be a client side caching issue on the OS not playing nicely with persona and sent me a KB to check out. We had some unrelated back end AD and network issues we had to resolve the past few days. But im currently monitoring our pool after following the kb steps to fully disable CSC. I dont have my work email in front of me so ill post the kb tomorrow when i get back to the grind.

[cdubz]

Here is the KB they referred me to:

View Persona Management features do not function when Windows Client-Side Caching is in effect (2016...

[eccl1213]

Thanks for the link.

I don't think that applies to our pools.  We already have CSC completely disabled.  The VMWare Optimization tool disables the service which we run as part of our image.

But we can confirm its a persona issue.  If we disable the persona GPO we have no issues.

[cdubz]

The OS Optimization does disable most of the CSC but it doesnt perform the registry key entry mentioned in the article. We run the optimization tool as well  Mostly irrelevant anyways as i had one machine crash this morning.

On the persona GPO side of things, what settings are you enabling or configuring on your GPO?  I too am on the believe that there is something within persona managment that does not play well with the new windows 10 OS and how they merged in the new start menu/metro app functionality.

[h3nkY]

This is interesting. You hit BSOD with View 7.0.

Can you let me know ticket number so I can take a look at it?

[eccl1213]

Ours is SR 16180652207

Our persona setup is pretty straight forward.  We redirect all the major folders (desktop, my docs etc).

We have local appdata redirect turned on but then excluded all except for one small application folder we need.

We do not preload any folders.

Browsing the persona folders doesn't show any unneeded folders syncing to appdata\local.  They look pretty clean.

[cdubz]

SR 16138597206. Enjoy

[amr12]

Had this same issue but with a little different set up, this fixed us up.

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=21172...

[cdubz]

Thanks for the heads up amr12.  We are actually running 6.0 update 2 with horizon 7.0.1 so this wouldnt be applicable to my case, but an informative read.

[h3nkY]

Please upload your memory dump file to support ticket so I can see it.

Thanks.

[h3nkY]

Please upload your memory dump to support ticket.

It will not be in agent log bundle. You need to take manually from c:\windows folder.