We are experiencing the same issue on a larger scale. We have approx 50 users on our pilot View program and everyone receives the following message throughout the day "Network Connection Lost" which is a message box popping up from the zeroclient. The message appears for a few seconds and then lets the user resume doing their work. If the message pops up for longer then a few seconds then the user is disconnected and kicked back out to the Zeroclient connection screen. Sometimes they can log right back in and resume work other times they may receive the message that no desktop source is available. At that point they have to wait a little while then log back in.
We once saw this happen to a user while we were remoting into their system via a third party remoting tool. On the remoting side the virtual desktop locks itself so you see the windows box saying that so and so user is logged in only that user or a administrator can unlock the system. Meanwhile the user is sent back to the zero client login screen with a message saying "Session Lost!"
We have a open SR with Vmware for the past 4 weeks and still have been unable to resolve the issue. Technicians believe it could be caused by a network problem. We have 1GB connections at least all the way from the zeroclient to the virtual desktop and all in the same building. Continous network monitoring never shows bandwidth or utilization ever maxing out.
If I were you least rule out basic networking at that user's port since you only have one user with the problem. Check the port make sure its configured to Auto if the port is running 1GB. Make sure the zeroclient is set to Auto. Change out network cable, Monitor Switch, Monitor users port etc.
Something for you to consider if you haven't already, we had a sister company that experienced a very similiar problem which in the end they resolved. Their disconnects were being caused by their HP SAN which showed the storage processors at very high utlization. If you haven't already, if you have a SAN, make sure to analyze that particluar user's datastore or whatever disk are hosting their Virtual machine. Monitor it to see if the disk latency is ever reaching 30MS. You can even use ESXTOP on the ESX host to monitor the LUN or disks to at least rule storage as a possible culprit.
Below are some of the configurations we are using, please let me know what you are using that is similiar to us, perhaps we have something in common that not alot of other users have and could be key to the problem. If we don't resolve this issue soon this may be the end of our View days.
ESXI 4 Update 1
Virtual Center 4 running on Windows Server 2008 R2 64bit
View Connection Server running on Windows Enterprise 2003 64bit
ESXI running on HP Blade Servers inside a HP Blade Enclousure
Cisco Switches end to end
Zero Clients are Dell FX100 running Teradici firmware 3.1.2
Storage is running on EMC Clarion SAN, OS running on Fiberchannel 15k Drivers Raid 10
Virtual Desktops running Windows XP SP3, Typical software Symantec Endpoint, Cisco Clean Access Agent
Sorry I hadn't gotten back to this thread in a while.. I've been away.
I can now confirm that this issue is happening on multiple machines, so it is not isolated to one desktop.
The affected machines are also in separate buildings on the campus, so it's not an issue with the network in the one building either.
I am using:
ESXi 4.01 U1
Vcenter 4 running on Windows 2003 R2 64bit
View connection server on Windows 2003 R2 32-bit
ESXi servers: 3 Sunfire 4150's and 1 Dell poweredge 2900
Storage: 2x EqualLogic PS6000E's w/ 1TB SATA disks (raid 50)
Zero clients: Samsung NC240 w/ 3.1.2 firmware
VM's are XP sp3 w/ latest view agent
I have the exact same issue with the same zero-clients.
The network is fine and this only happens with the zero-clients.
Our software-based PCoIP clients works just fine and they are on the exact same network and on the same vmware-environment.
So as many others have concluded i'm sure the network is just fine and is not causing this problem, there must be something with the teradici-fw, at least something in their implementation.
I just looked thru the logs on one of our samsungs and noticed that our lab-wide security scanner is hammering the PCOIP monitor right before a disconnect..
At leasat that is what it looks like to me.
Maybe this is my issue. Is anyone else with this issue doing security scans on their network?
i am not sure if you solved this yet or not but..
This might be off topic sort of. we are testing the p20 and the
nc240.. both of them get disconnected after a peroid of idle.and this
has nothing to do with the screen saver.
However we noticed that moving the guest OS to a different OU (one without the production user settings) solved the disconnection issue. And yes there would be times were it would say there are no desktop resources availible. but since moving that guest to a different OU we have not had a disconnect yet.
I am tentatively calling this issue resolved.
I got a unreleased patch from teradici that seems to have fixed the disconnect issue for me.
So far it is looking good...
We are having the same issue with the Samsung NC240's we are using as a test pilot and this is one issue stopping me from moving forward with a deployment. Do you know if this patch will be released soon or should I contact Teradici?
I"m not sure if hes referring to the latest release by teradici but they just released firmware 3.2 this last friday. If you haven't already try 3.2. We are still having the "Network connection lost" messages but haven't done any testing with 3.2 yet. Will try soon though.
You should contact teradici.
The patch was not to their firmware, but to the pcoip server execuatable installed along with the view agent.
We have tried using the new PCOIP executable which you are talking about that comes with View 4.5 agent and still receive the message. Can you verify if the patch exe that you are using is newer then this one? If so that means teradici came up with some fixes post View 4.5. Here is the 4.5 RC1 one
DateModified: July 16 2010
This execuatable is from June 18 2010. 1.23MB
I will double check with my users and see how things are going.
Maybe I finally have them trained and they just stopped complaining....
To tie off this post, VMware View 4.5 and 4.0.2 resolved a number of session disconnects when connecting via PCoIP Zero Clients.
The "Network Connection Lost" message on PCoIP Zero Clients indicates that the information on the display may be stale due to a momentary network disruption or a network disconnection. This message is triggered quickly for real-time sensitive applications (ie investment bank trading floors) to ensure business decisions are not made on stale display data. This message should be disabled for mainstream virtual desktops. For more details see http://techsupport.teradici.com/ics/support/default.asp?deptID=15164&task=knowledge&questionID=82
If you are experiencing session disconnects with PCoIP Zero Clients with firmware 3.1.x or 3.2.x and VMware View 4.5, please check that you have disabled SLP discovery in the PCoIP Zero Client. For more details see http://techsupport.teradici.com/ics/support/default.asp?deptID=15164&task=knowledge&questionID=498 (self-registration may be required).
Please submit a ticket at techsupport.teradici.com if you have further questions.
- Teradici Systems Engineering