Users unexpectedly disconnected from VD - Machine unavailable.

[MickQuick]

Hello,

i've just set up a new View 4 infrastructure, i've created an automatic persistent Pool managed by Composer. Virtual Infrastructure is based on vSphere 4 Enterprise and vCenter. Desktop Pool is composed by 10 desktop and 5/6 users that simultaneously connect to their desktop. Connection is made by vmware view client, PCoIP protocol. Agent and client version is the same.

I've tested VDI for weeks without any issue, now users have just start to use it and they are notifyng me that they are unexpectedly disconnected from their windows XP virtual desktop.

On connection manager server's Event i've found this errors

Type: Warn

Time: 04/05/10 17.59.45

Module: DesktopsHandler

Thread: TP-Processor3

Details : (SESSION:C6193D0EC3A1D2E6AA0D09E528A82291;CAD487ADA9C7FCE7B2D3E7B1B39DEFD2) Pool control for desktop XXXXXVDIXXX reports machine assigned to DOMAIN\ACCOUNTAD is unavailable

and the other infos about disconnection.

Since now I haven't tried anything ( i need PCoIP for Nortel SoftPhone , RDP is useless ).

Any idea?

Thanks

Michele

[NJaure]

Hi, we've got a similar issue. Please check if the View Connection Server loses network connectivity. If does, it is the same problem as ours.

Our View Server gets disconnected form network, so the users are disconnected form their VD. After restarting the service, network gets back.

[MickQuick]

Hello,

it is not a network problem, i looked at connection manager event log and found

***********

Warning

event id :104

(29B5FBB80D710BE6A34C82AD9439FD93) Conveyor 0 IOException. Remote address is null: An existing connection was forcibly closed by the remote host

and

(29B5FBB80D710BE6A34C82AD9439FD93) User urxd183 Secure Gateway Server session ended - session ID: 29B5FBB80D710BE6A34C82AD9439FD93

***********

I searched on forums and found that could be related to Video Memory that should be set to an integer number ( 32 / 64 mb ) or to "automatic". I did that but issue persists, i then increased total memory of the VM , i'm waiting for users feedback.

I checked how users are disconnected: remote session freezes, no interaction is available with desktop and after few seconds view client shuts down.

Thanks

Michele

[krowczynski]

Please check out if it is a session timeout in the View Connection Server.

MCP, VCP

[MickQuick]

Also memory increase didn't work.

These are global settings

Session timeout: 600 minutes

Use SSL for client connections: Yes

Reauthenticate Secure VPN connections after network interruption: Yes

Message security mode: Disabled

Direct connection for Offline Desktop operations: No

Use SSL for Offline Desktop operations: No

Disable Single Sign-On for Offline Desktop operations: No

Pre-login message: No

Display warning before forced logoff: Yes

Are there any other settings?

Could be related to the computer that my colleagues use to start view client? It's an HP 5750 windows XP SP3 with Sophos Antivirus and many other applications, in fact i use "real" computer for testing VDI. If we go in production i think we'll buy HP thin clients.

Thanks

Michele

[MickQuick]

Hello,

problem seems to be client related. I gave to an user an old HP thin client, he used it for 3 days and he noticed disconnection only 2 times ( he notified me also slowness using nortel softphone ). I then set up a new Pc with a clean installation of windows xp SP3 + critical updates , drivers and view client, Pc is a HP 5750. As he started to use it disconnections happened every 5/10 minutes, it's impossibile to work.

Disconnections happens randomly, if you work, if you let VD iconized on taskbar, really don't know how to go on.

Michele

[onefstprld]

There is a good way to test the reliability of your connection server, there is a checkbox on the global settings to allow "direct connection" to your View desktops. If you enable this it bypasses the SSL tunnel of the connection server and the View client connects directly to the VM. If the drops stop occurring then its the Connection Server, if drops are still happening it could be a client network connection issue.

[JugOfTuiPlease]

I just had a huge disruption like this as weel this morning, witht he same event log error.

I already use direct connection for the VMs.

I'm stumped to what caused this.

[JugOfTuiPlease]

My case looks like it was overpopulating my vswitch with VMs.

I increased the number of ports on the vswitch and am now holding my breath as users are now starting to connect

[zackangelo]

Did this solve your problem?

I'm experiencing identical symptoms.

[JugOfTuiPlease]

I ended up having to reboot the hosts, first vmotioning desktops around to minimise downtime.

This resolved it, but I'm still in the dark wrt the cause.