Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

36224

November 13th, 2010 11:00

2 minute delay on opening .pit file (from DMZ / SSL gateway)

From the beginning we have this strange issue:

Customers with Internet Explorer report that the vworkspace web interface is extreme slow.

When they login and click on their virtual desktop, it starts loading something, but nothing is happening.

Normally you would see a popup screen telling the user that the desktop is starting up or something.

It almost looks like Internet Explorer is waiting to get data and then after 2 minutes it finally starts doing something.

We have not seen this behavor with Firefox. It only happens with IE and mainly in combination with Vista/Windows 7.

We had a support ticket open for months but the support guys couldn't reproduce it or didn't know what to do.

However people keep complaining "This VDI is very slow" ... they mean the startup.

I have included a video so you can see exactly what I mean. If nobody in this community knows what this is, then whats next??

vworkspace video capture

Message was edited by: René Kemp Changed topic title

17 Posts

June 30th, 2011 14:00

After a great deal of debugging/troubleshooting, we have narrowed this issue down to an IIS7+ issue currently with the installation of Web Access.

Currently we create an Application Pool to run Web Access and set the 'Managed Pipeline' to 'Integrated'.

Switching this 'Managed Pipeline' to 'Classic' mode appears to resolve the issue.

Quest Support also has a Knowledge Base article on this issue now.

48 Posts

November 14th, 2010 16:00

Hi Renee,

The first time a site is hit IIS must load the asp.net worker process and basically start up the .NET runtime. That takes a little bit of time and is very noticeable on first hit. After a period of inactivity the asp.net application is shut down and then has to be restarted.

You can speed up the startup by disabling the code signing lookup.

Edit c:\windows\Microsoft.NET\Framework\v2.0.50727\aspnet.config and add the text shown in bold



   
       
       
       
       
      
   

The other thing to do is to increase the asp.net worker process idle timeout with the following command:

cscript c:\Inetpub\AdminScripts\adsutil.vbs set W3SVC/AppPools/DefaultAppPool/IdleTimeout 300

This increase the idle timeout to 300 minutes (from the default of 20) and during the day will help the web portal load much more quickly.

Restart IIS and things should be somewhat faster.

IE is pretty anal about SSL certificates, particulalrly IE 7 and 8, so another one of the things that may help is to try unchecking the server and publisher certificate revocation options (tools > internet options > advanced (Security section)) on the client machine.

Hopefully that will speed things up a bit for you.

regards,

Rick

74 Posts

November 15th, 2010 06:00

Rick Mack wrote:

The first time a site is hit IIS must load the asp.net worker process and basically start up the .NET runtime. That takes a little bit of time and is very noticeable on first hit. After a period of inactivity the asp.net application is shut down and then has to be restarted.

This is a different problem we had, but indeed by adjusting some IIS settings can be solved. What we did was setting the Idle Timeout to 0 and the Regualr Time Interval (Recycling) to 0.

The problem we have now happens only with Internet Explorer. The web-interface is ok, refreshing works fine. It goes wrong at the point when you click on the virtual desktop icon. I don't think it has anything to do with caching/worker processes because it never happens with firefox.

IE is pretty anal about SSL certificates, particulalrly IE 7 and 8, so another one of the things that may help is to try unchecking the server and publisher certificate revocation options (tools > internet options > advanced (Security section)) on the client machine.

Very good suggestion, this was one of the first things we tried (and didn't work ). Still, it must have something to do with SSL.

74 Posts

November 18th, 2010 07:00

Maybe one of the developers can clarify some things?

What is happening at the moment you press the virtual desktop icon in the web interface, just after the moment the 'Resource is loading ...' popup, but before the broker is sending a powerup signal to vmware?

Between this period, we get 2 minute timeout in Internet Explorer.

It only seems to happen from the outside, when connection through SSL. At this point, only connections are allowed through port 443. It almost looks like Internet Explorer is waiting for a connection that can't be made. If we try this on the internal network, it works without delays, but this might be because you can access all servers.

74 Posts

November 26th, 2010 10:00

Some more information:

Everytime a user experiences this weird 120 second timeout in internet explorer, we get the following error in the eventviewer. Any idea's?

PLEASE?

 
 
 
 
 
Log Name:      Application
Source:        ASP.NET 2.0.50727.0
Date:          11/26/2010 1:33:08 PM
Event ID:      1309
Task Category: Web Event
Level:         Warning
Keywords:      Classic
User:          N/A
Computer:      SERVERX
Description:
Event code: 3001
Event message: The request has been aborted.
Event time: 11/26/2010 1:33:08 PM
Event time (UTC): 11/26/2010 12:33:08 PM
Event ID: 5b9ecfd8780f4ec7add768ab35e1a0c3
Event sequence: 60
Event occurrence: 1
Event detail code: 0
Application information:
    Application domain: /LM/W3SVC/1/ROOT/Provision-12-129352482218501681
    Trust level: Full
    Application Virtual Path: /Provision
    Application Path: C:\inetpub\wwwroot\Provision\
    Machine name: SERVERX
Process information:
    Process ID: 3788
    Process name: w3wp.exe
    Account name: SERVERX\QWAUser1
Exception information:
    Exception type: HttpException
    Exception message: Request timed out.
Request information:
    Request path: /Provision/web-it/default.aspx
    User host address: xxx.xxx.xxx.xxx
    User:
    Is authenticated: False
    Authentication Type:
    Thread account name: SERVERX\QWAUser1
Thread information:
    Thread ID: 14
    Thread account name: SERVERX\QWAUser1
    Is impersonating: False
    Stack trace:
Custom event details:  Event Xml:
http://schemas.microsoft.com/win/2004/08/events/event ">
 
   
    1309
    3
    3
    0x80000000000000
   
    14558
    Application
    SERVERX
   
 
 
    3001
    The request has been aborted.
    11/26/2010 1:33:08 PM
    11/26/2010 12:33:08 PM
    5b9ecfd8780f4ec7add768ab35e1a0c3
    60
    1
    0
    /LM/W3SVC/1/ROOT/Provision-12-129352482218501681
    Full
    /Provision
    C:\inetpub\wwwroot\Provision\
    SERVERX
   
   
    3788
    w3wp.exe
    SERVERX\QWAUser1
    HttpException
    Request timed out.
    /Provision/web-it/default.aspx
    xxx.xxx.xxx.xxx
   
   
    False
   
   
    SERVERX\QWAUser1
    14
    SERVERX\QWAUser1
    False
   
   
 

173 Posts

December 1st, 2010 09:00

It would be good to see the screenshot of your WA configuration for the Secure Gateway but I realise this might be confidential information.

74 Posts

December 2nd, 2010 08:00

Sure, anything to solve this weird problem:

Clip1.png

(externaldomain.com is just an example)

173 Posts

December 4th, 2010 09:00

Do you front-end  WA with the Quest Secure Gateway?

74 Posts

December 6th, 2010 13:00

Yes, they are installed on the same server. WA and RDP go thru quest's SSL gateway.

48 Posts

December 6th, 2010 21:00

Hi Renee,

Firstly, I wouldn't use the SSL Gateway for internal users unless you recognize that it will inject a few milliseconds latency and has practical throughput limits. You mentioned earlier that you'd like to know what happens when you press the desktop icon. The diagram below might help:

wf.jpg

I can provide the visio file if needed.

Process Monitor from http://live.sysinternals.com (procmon.exe, set filter, processname is pntsc.exe) will also help pinpoint where the delays are happening. Look at the event timestamps to find long gaps where nothing is happening.

regards,

Rick

74 Posts

December 7th, 2010 17:00

Hello Rick,

Internal users use another server without SSL. They don't experience the long delay I'm describing.

Looking at the diagram, it is exacly between 'User clicks an application icon for launching' and 'Web-IT sends application launc request to Broker' that randomly takes a very long time, only in Internet Explorer.

So when someone click on the icon, there is a 2 minute timeout, the broker is doing nothing. Then suddenly after 2 minutes the broker begins to do something.

48 Posts

December 7th, 2010 20:00

Hi Rene,

If the delay is due to something like a name resolution problem, then I would expect that logons would also be very slow since the connection broker is used for authentication and application launch operations. In that case I would be tempted to add the connection broker(s) to the hosts file on the web portal IIS server, and add the web portal into the connection broker(s) hosts file.

Are you using NLB between 2 web portals? That can make life interesting if DNS isn't quite right

I'd also suggest you remove the internal users web access URL entry in your SSL/VPN configuration since it could confuse things for you.

regards,

Rick

74 Posts

December 21st, 2010 07:00

We are using IP's, so it shouldn't be a DNS problem. Besides, isn't it strange that it only happens with IE and not Firefox?

This looks more like a problem with the web application, which is running in IE7 compatibility mode (IE=EmulateIE7).

We are not using NLB.

I'd also suggest you remove the internal users web access URL entry in your SSL/VPN configuration since it could confuse things for you.

Documentation says both fields MUST be filled in. I tried it with the internal URL removed anyeway, same result.

173 Posts

December 21st, 2010 08:00

Just as a test, can you do me a favor and disable the highlighted check in IE (see att)

1 Message

December 23rd, 2010 23:00

This is a known issue with the .Net Framework 1.1, 2.0, 3.0 and 3.5 You have a few options. Upgrade to the 4.0 Framework, preferably with IIS 7 which has the ability to ping your site and deep it alive. If you are not able to upgrade you could also use a ping utility to keep your site "alive". Basically what this does is it hits your site every few minutes so that you do not encounter the 20 min. default timeout period of you application domain. When you application domain times out it is reloaded on the next request, hence the slowness you are experiencing.

Find a "Ping/Keep Alive" service here

No Events found!

Top