Underlying connection was closed

[Rajeev Tipnis]

http://support.microsoft.com/default.aspx?scid=kb;en-
us;819450

Questions:
1) Is this patch (fix) applicable to the .NET 1.1
Framework as well? That is, if we have Framework 1.1 (On
Win2K3/IIS6.0), do we still need this patch?

2) Does it specifically solve only this error:
The underlying connection was closed. An unexpected error
occured on send.
and not these (which is what we are seeing):

- The underlying connection was closed. Unable to connect
to the remote server. OR
- The underlying connection was closed. The proxy name
could not be resolved, verify correct proxy configuration.

3) Would this not happen when the webservices client is
not an ASP.NET client?

Thanks
Rajeev

[Daniel Roth [MSFT]]

Hi Rajeev,

I am in the process of investigating your questions. I will get back to you as soon as I have an answer.

Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
Subject: Underlying connection was closed
Date: Thu, 19 Aug 2004 14:59:51 -0700
Lines: 23
Message-ID: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
Thread-Index: AcSGN9pVbhl3R/z4Q7iWwoav0NPNX g==
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6290
NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices

http://support.microsoft.com/default.aspx?scid=kb;en-
us;819450

Questions:
1) Is this patch (fix) applicable to the .NET 1.1
Framework as well? That is, if we have Framework 1.1 (On
Win2K3/IIS6.0), do we still need this patch?

2) Does it specifically solve only this error:
The underlying connection was closed. An unexpected error
occured on send.
and not these (which is what we are seeing):

- The underlying connection was closed. Unable to connect
to the remote server. OR
- The underlying connection was closed. The proxy name
could not be resolved, verify correct proxy configuration.

3) Would this not happen when the webservices client is
not an ASP.NET client?

Thanks
Rajeev

--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all responses to this message are best directed to the newsgroup/thread from which they originated.

[Daniel Roth [MSFT]]

Hi Rajeev,

I've answered your questions below:

1) I believe this patch does apply to .Net 1.1.
2) No, it does not solve the other cases since they are mostly a result of user configuration problems.
3) Yes, the fix does apply when the client is not an ASP.NET client.

For the issues you are having, you should verify that your proxy settings are correct and that you are correctly specifying a valid server. Quite possibly these two issues
are related. If you are still having these problems, I might be able to help if you provide some more details.

Let me know if you have further questions.

Daniel Roth
--------------------
X-Tomcat-ID: 809948391
References: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----=_NextPart_0001 _6ADD76EC"
Content-Transfer-Encoding: 7bit
From: da****@online.m icrosoft.com (Daniel Roth [MSFT])
Organization: Microsoft
Date: Thu, 26 Aug 2004 23:37:23 GMT
Subject: RE: Underlying connection was closed
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices
Message-ID: <np************ **@cpmsftngxa10 .phx.gbl>
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Lines: 120
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6375
NNTP-Posting-Host: TOMCATIMPORT1 10.201.218.122

Hi Rajeev,

I am in the process of investigating your questions. I will get back to you as soon as I have an answer.

Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
Subject: Underlying connection was closed
Date: Thu, 19 Aug 2004 14:59:51 -0700
Lines: 23
Message-ID: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
Thread-Index: AcSGN9pVbhl3R/z4Q7iWwoav0NPNX g==
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6290
NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices

http://support.microsoft.com/default.aspx?scid=kb;en-
us;819450

Questions:
1) Is this patch (fix) applicable to the .NET 1.1
Framework as well? That is, if we have Framework 1.1 (On
Win2K3/IIS6.0), do we still need this patch?

2) Does it specifically solve only this error:
The underlying connection was closed. An unexpected error
occured on send.
and not these (which is what we are seeing):

- The underlying connection was closed. Unable to connect
to the remote server. OR
- The underlying connection was closed. The proxy name
could not be resolved, verify correct proxy configuration.

3) Would this not happen when the webservices client is
not an ASP.NET client?

Thanks
Rajeev

--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all responses to this message are best directed to the newsgroup/thread from which they originated.

--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all responses to this message are best directed to the newsgroup/thread from which they originated.

[Rajeev Tipnis]

Hi Daniel -
Thanks for looking into this.

When you say, this patch applied to 1.1, do you mean we
still need to get this patch (if we already have 1.1)?
In any case, we are in the process of getting this patch.

However, what I dont understand is - how can the
configuration be incorrect as this happens
only "intermittently ".

So, in my load test case, I have a load test application
that runs against a web service. When the number of
simultaneous connections go over 20-25 range to the web
service, we begin to see these errors.

In my case the scenario is as follows:
Two separate machines:
1. Web Services Client (for load testing):
A simple .net console app (.NET 1.1 Framework)
2. Web Services Server:
A .Net web service hosted on IIS 6.0 (on Win2K3)
on .NET 1.1 Framework.

I start to see these errors when the load goes
beyond a certain threshold (about 25 connections, I
believe are open at that point).

What I have experimented with so far:
- Cleared the keep-alive flag on the IIS and on the
client by overriding the client proxy methods
- IIS 6.0 is configured to receive unlimited connections
and configured to have really long timeouts

If the configuration was incorrect, I would expect that
this would fail each time, but it fails with these errors
only under load.

Thanks in advance.

Rajeev

-----Original Message-----
Hi Rajeev,

I've answered your questions below:

1) I believe this patch does apply to .Net 1.1.
2) No, it does not solve the other cases since they are mostly a result of user configuration problems.3) Yes, the fix does apply when the client is not an ASP.NET client.
For the issues you are having, you should verify that your proxy settings are correct and that you are
correctly specifying a valid server. Quite possibly
these two issuesare related. If you are still having these problems, I might be able to help if you provide some more details.
Let me know if you have further questions.

Daniel Roth
--------------------
X-Tomcat-ID: 809948391
References: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="---- =_NextPart_0001 _6ADD76EC"Content-Transfer-Encoding: 7bit
From: da****@online.m icrosoft.com (Daniel Roth [MSFT])
Organization : Microsoft
Date: Thu, 26 Aug 2004 23:37:23 GMT
Subject: RE: Underlying connection was closed
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservicesMessage-ID: <np************ **@cpmsftngxa10 .phx.gbl>
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Lines: 120
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6375NNTP-Posting-Host: TOMCATIMPORT1 10.201.218.122

Hi Rajeev,

I am in the process of investigating your questions. I will get back to you as soon as I have an answer.
Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
Subject: Underlying connection was closed
Date: Thu, 19 Aug 2004 14:59:51 -0700
Lines: 23
Message-ID: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
Thread-Index: AcSGN9pVbhl3R/z4Q7iWwoav0NPNX g==
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6290NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices
http://support.microsoft.com/default.aspx?scid=kb;en-
us;819450

Questions:
1) Is this patch (fix) applicable to the .NET 1.1
Framework as well? That is, if we have Framework 1.1 (On
Win2K3/IIS6.0), do we still need this patch?

2) Does it specifically solve only this error:
The underlying connection was closed. An unexpected erroroccured on send.
and not these (which is what we are seeing):

- The underlying connection was closed. Unable to connectto the remote server. OR
- The underlying connection was closed. The proxy name
could not be resolved, verify correct proxy configuration.
3) Would this not happen when the webservices client is
not an ASP.NET client?

Thanks
Rajeev

--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are
subject to the terms specified athttp://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all responses to this message are best directed to the
newsgroup/thread from which they originated.
--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are
subject to the terms specified athttp://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all

responses to this message are best directed to the
newsgroup/thread from which they originated.

[Rajeev Tipnis]

Today, I seem to be seeing these errors instead (so the
same underlying connection errors - but says "The request
was canceled" and in the following alternating sequence:

Exception: System.Net.WebE xception: The operation has
timed-out.
at System.Net.Http WebRequest.GetR equestStream()
at
System.Web.Serv ices.Protocols. SoapHttpClientP rotocol.Invok
e(String methodN
ame, Object[] parameters)

followed by:

Exception: System.Net.WebE xception: The underlying
connection was closed: The request was canceled.
at
System.Web.Serv ices.Protocols. WebClientProtoc ol.GetWebResp
onse(WebRequest request)
at
System.Web.Serv ices.Protocols. HttpWebClientPr otocol.GetWeb
Response(WebReq uest request)
at
System.Web.Serv ices.Protocols. SoapHttpClientP rotocol.Invok
e(String methodName, Object[] parameters)

Again, does not happen all the time - I would say about
35% of the times under load.

Thx

[Daniel Roth [MSFT]]

You might be getting the "The underlying connection was closed. Unable to connect to the remote server" error because
too many connections are in the TIME_WAIT state. When this happens the system may run out of the random port range.
Clearing the KeepAlive flag might actually be aggravating this situation.

Please try entering "netstat -p TCP" at the command prompt and see how many TIME_WAIT connections are found.

Also, do you have code that we could use to reproduce these errors for further investigation?

Thank you for your patience.

Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
References: <99************ *************** *@phx.gbl> <np************ **@cpmsftngxa10 .phx.gbl>
<BI************ **@cpmsftngxa10 .phx.gbl>
Subject: RE: Underlying connection was closed
Date: Fri, 27 Aug 2004 12:41:42 -0700
Lines: 170
Message-ID: <17************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
thread-index: AcSMbeDGoXHBF+5 uQYaTicjNqvUzdQ ==
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6386
NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices
Hi Daniel -
Thanks for looking into this.

When you say, this patch applied to 1.1, do you mean we
still need to get this patch (if we already have 1.1)?
In any case, we are in the process of getting this patch.

However, what I dont understand is - how can the
configuration be incorrect as this happens
only "intermittently ".

So, in my load test case, I have a load test application
that runs against a web service. When the number of
simultaneous connections go over 20-25 range to the web
service, we begin to see these errors.

In my case the scenario is as follows:
Two separate machines:
1. Web Services Client (for load testing):
A simple .net console app (.NET 1.1 Framework)
2. Web Services Server:
A .Net web service hosted on IIS 6.0 (on Win2K3)
on .NET 1.1 Framework.

I start to see these errors when the load goes
beyond a certain threshold (about 25 connections, I
believe are open at that point).

What I have experimented with so far:
- Cleared the keep-alive flag on the IIS and on the
client by overriding the client proxy methods
- IIS 6.0 is configured to receive unlimited connections
and configured to have really long timeouts

If the configuration was incorrect, I would expect that
this would fail each time, but it fails with these errors
only under load.

Thanks in advance.

Rajeev

-----Original Message-----
Hi Rajeev,

I've answered your questions below:

1) I believe this patch does apply to .Net 1.1.
2) No, it does not solve the other cases since they are mostly a result of user configuration problems.3) Yes, the fix does apply when the client is not an ASP.NET client.
For the issues you are having, you should verify that your proxy settings are correct and that you are
correctly specifying a valid server. Quite possibly
these two issuesare related. If you are still having these problems, I might be able to help if you provide some more details.
Let me know if you have further questions.

Daniel Roth
--------------------
X-Tomcat-ID: 809948391
References: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="---- =_NextPart_0001 _6ADD76EC"Content-Transfer-Encoding: 7bit
From: da****@online.m icrosoft.com (Daniel Roth [MSFT])
Organization : Microsoft
Date: Thu, 26 Aug 2004 23:37:23 GMT
Subject: RE: Underlying connection was closed
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservicesMessage-ID: <np************ **@cpmsftngxa10 .phx.gbl>
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Lines: 120
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6375NNTP-Posting-Host: TOMCATIMPORT1 10.201.218.122

Hi Rajeev,

I am in the process of investigating your questions. I will get back to you as soon as I have an answer.
Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
Subject: Underlying connection was closed
Date: Thu, 19 Aug 2004 14:59:51 -0700
Lines: 23
Message-ID: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
Thread-Index: AcSGN9pVbhl3R/z4Q7iWwoav0NPNX g==
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6290NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices
http://support.microsoft.com/default.aspx?scid=kb;en-
us;819450

Questions:
1) Is this patch (fix) applicable to the .NET 1.1
Framework as well? That is, if we have Framework 1.1 (On
Win2K3/IIS6.0), do we still need this patch?

2) Does it specifically solve only this error:
The underlying connection was closed. An unexpected erroroccured on send.
and not these (which is what we are seeing):

- The underlying connection was closed. Unable to connectto the remote server. OR
- The underlying connection was closed. The proxy name
could not be resolved, verify correct proxy configuration.
3) Would this not happen when the webservices client is
not an ASP.NET client?

Thanks
Rajeev

--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are
subject to the terms specified athttp://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all responses to this message are best directed to the
newsgroup/thread from which they originated.
--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are
subject to the terms specified athttp://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all

responses to this message are best directed to the
newsgroup/thread from which they originated.

--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are subject to the
terms specified at http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all responses to this message are best directed to the newsgroup/thread
from which they originated.

[Daniel Roth [MSFT]]

Hi Rajeev,

This behavior may be due to a known issue which requires a service patch (Hotfix #3028) which you can request through
Microsoft Product Support Services. The issue occurs when you are sending more requests then can be handled, and the
requests are queued internally. If these queued requests timeout before even hitting the wire, you may encounter the
issue fixed in the service patch.

Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
References: <99************ *************** *@phx.gbl> <np************ **@cpmsftngxa10 .phx.gbl>
<BI************ **@cpmsftngxa10 .phx.gbl> <17************ *************** *@phx.gbl>
Subject: RE: Underlying connection was closed
Date: Fri, 27 Aug 2004 13:55:58 -0700
Lines: 33
Message-ID: <02************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
Thread-Index: AcSMeEDgyaPgu+L KR++naw+Hixg/jA==
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6387
NNTP-Posting-Host: tk2msftngxa12.p hx.gbl 10.40.1.164
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices
Today, I seem to be seeing these errors instead (so the
same underlying connection errors - but says "The request
was canceled" and in the following alternating sequence:

Exception: System.Net.WebE xception: The operation has
timed-out.
at System.Net.Http WebRequest.GetR equestStream()
at
System.Web.Serv ices.Protocols. SoapHttpClientP rotocol.Invok
e(String methodN
ame, Object[] parameters)

followed by:

Exception: System.Net.WebE xception: The underlying
connection was closed: The request was canceled.
at
System.Web.Serv ices.Protocols. WebClientProtoc ol.GetWebResp
onse(WebRequest request)
at
System.Web.Serv ices.Protocols. HttpWebClientPr otocol.GetWeb
Response(WebReq uest request)
at
System.Web.Serv ices.Protocols. SoapHttpClientP rotocol.Invok
e(String methodName, Object[] parameters)

Again, does not happen all the time - I would say about
35% of the times under load.

Thx

--

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are subject to the
terms specified at http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all responses to this message are best directed to the newsgroup/thread
from which they originated.

[AShah]

Did you get a solution to this issue ? I am seeing the exactly the
same behaviour. When ever I load my application with many requests I
start seeing the same errors.

Thanks.
da****@online.m icrosoft.com (Daniel Roth [MSFT]) wrote in message news:<T3******* *******@cpmsftn gxa10.phx.gbl>. ..

You might be getting the "The underlying connection was closed. Unable to connect to the remote server" error because
too many connections are in the TIME_WAIT state. When this happens the system may run out of the random port range.
Clearing the KeepAlive flag might actually be aggravating this situation.

Please try entering "netstat -p TCP" at the command prompt and see how many TIME_WAIT connections are found.

Also, do you have code that we could use to reproduce these errors for further investigation?

Thank you for your patience.

Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
References: <99************ *************** *@phx.gbl> <np************ **@cpmsftngxa10 .phx.gbl>
<BI************ **@cpmsftngxa10 .phx.gbl>
Subject: RE: Underlying connection was closed
Date: Fri, 27 Aug 2004 12:41:42 -0700
Lines: 170
Message-ID: <17************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
thread-index: AcSMbeDGoXHBF+5 uQYaTicjNqvUzdQ ==
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6386
NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices
Hi Daniel -
Thanks for looking into this.

When you say, this patch applied to 1.1, do you mean we
still need to get this patch (if we already have 1.1)?
In any case, we are in the process of getting this patch.

However, what I dont understand is - how can the
configuration be incorrect as this happens
only "intermittently ".

So, in my load test case, I have a load test application
that runs against a web service. When the number of
simultaneous connections go over 20-25 range to the web
service, we begin to see these errors.

In my case the scenario is as follows:
Two separate machines:
1. Web Services Client (for load testing):
A simple .net console app (.NET 1.1 Framework)
2. Web Services Server:
A .Net web service hosted on IIS 6.0 (on Win2K3)
on .NET 1.1 Framework.

I start to see these errors when the load goes
beyond a certain threshold (about 25 connections, I
believe are open at that point).

What I have experimented with so far:
- Cleared the keep-alive flag on the IIS and on the
client by overriding the client proxy methods
- IIS 6.0 is configured to receive unlimited connections
and configured to have really long timeouts

If the configuration was incorrect, I would expect that
this would fail each time, but it fails with these errors
only under load.

Thanks in advance.

Rajeev

-----Original Message-----
Hi Rajeev,

I've answered your questions below:

1) I believe this patch does apply to .Net 1.1.
2) No, it does not solve the other cases since they are

mostly a result of user configuration problems.

3) Yes, the fix does apply when the client is not an

ASP.NET client.

For the issues you are having, you should verify that

your proxy settings are correct and that you are
correctly specifying a valid server. Quite possibly
these two issues

are related. If you are still having these problems, I

might be able to help if you provide some more details.

Let me know if you have further questions.

Daniel Roth
--------------------
X-Tomcat-ID: 809948391
References: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----

=_NextPart_0001 _6ADD76EC"

Content-Transfer-Encoding: 7bit
From: da****@online.m icrosoft.com (Daniel Roth [MSFT])
Organization : Microsoft
Date: Thu, 26 Aug 2004 23:37:23 GMT
Subject: RE: Underlying connection was closed
X-Tomcat-NG:

microsoft.publi c.dotnet.framew ork.webservices

Message-ID: <np************ **@cpmsftngxa10 .phx.gbl>
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Lines: 120
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl

microsoft.publi c.dotnet.framew ork.webservices :6375

NNTP-Posting-Host: TOMCATIMPORT1 10.201.218.122

Hi Rajeev,

I am in the process of investigating your questions. I

will get back to you as soon as I have an answer.

Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
Subject: Underlying connection was closed
Date: Thu, 19 Aug 2004 14:59:51 -0700
Lines: 23
Message-ID: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
Thread-Index: AcSGN9pVbhl3R/z4Q7iWwoav0NPNX g==
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl

microsoft.publi c.dotnet.framew ork.webservices :6290

NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG:

microsoft.publi c.dotnet.framew ork.webservices

http://support.microsoft.com/default.aspx?scid=kb;en-
us;819450

Questions:
1) Is this patch (fix) applicable to the .NET 1.1
Framework as well? That is, if we have Framework 1.1 (On
Win2K3/IIS6.0), do we still need this patch?

2) Does it specifically solve only this error:
The underlying connection was closed. An unexpected

error

occured on send.
and not these (which is what we are seeing):

- The underlying connection was closed. Unable to

connect

to the remote server. OR
- The underlying connection was closed. The proxy name
could not be resolved, verify correct proxy

configuration.

3) Would this not happen when the webservices client is
not an ASP.NET client?

Thanks
Rajeev

--

This posting is provided "AS IS" with no warranties, and

confers no rights. Use of included script samples are
subject to the terms specified at

http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all

responses to this message are best directed to the
newsgroup/thread from which they originated.

--

This posting is provided "AS IS" with no warranties, and

confers no rights. Use of included script samples are
subject to the terms specified at

http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all

responses to this message are best directed to the
newsgroup/thread from which they originated.

[AShah]

Did you get solution to this problem ? I am seeing exactly the same
problem when I load my system with many HTTP requests.

da****@online.m icrosoft.com (Daniel Roth [MSFT]) wrote in message news:<T3******* *******@cpmsftn gxa10.phx.gbl>. ..

You might be getting the "The underlying connection was closed. Unable to connect to the remote server" error because
too many connections are in the TIME_WAIT state. When this happens the system may run out of the random port range.
Clearing the KeepAlive flag might actually be aggravating this situation.

Please try entering "netstat -p TCP" at the command prompt and see how many TIME_WAIT connections are found.

Also, do you have code that we could use to reproduce these errors for further investigation?

Thank you for your patience.

Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
References: <99************ *************** *@phx.gbl> <np************ **@cpmsftngxa10 .phx.gbl>
<BI************ **@cpmsftngxa10 .phx.gbl>
Subject: RE: Underlying connection was closed
Date: Fri, 27 Aug 2004 12:41:42 -0700
Lines: 170
Message-ID: <17************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
thread-index: AcSMbeDGoXHBF+5 uQYaTicjNqvUzdQ ==
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl microsoft.publi c.dotnet.framew ork.webservices :6386
NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG: microsoft.publi c.dotnet.framew ork.webservices
Hi Daniel -
Thanks for looking into this.

When you say, this patch applied to 1.1, do you mean we
still need to get this patch (if we already have 1.1)?
In any case, we are in the process of getting this patch.

However, what I dont understand is - how can the
configuration be incorrect as this happens
only "intermittently ".

So, in my load test case, I have a load test application
that runs against a web service. When the number of
simultaneous connections go over 20-25 range to the web
service, we begin to see these errors.

In my case the scenario is as follows:
Two separate machines:
1. Web Services Client (for load testing):
A simple .net console app (.NET 1.1 Framework)
2. Web Services Server:
A .Net web service hosted on IIS 6.0 (on Win2K3)
on .NET 1.1 Framework.

I start to see these errors when the load goes
beyond a certain threshold (about 25 connections, I
believe are open at that point).

What I have experimented with so far:
- Cleared the keep-alive flag on the IIS and on the
client by overriding the client proxy methods
- IIS 6.0 is configured to receive unlimited connections
and configured to have really long timeouts

If the configuration was incorrect, I would expect that
this would fail each time, but it fails with these errors
only under load.

Thanks in advance.

Rajeev

-----Original Message-----
Hi Rajeev,

I've answered your questions below:

1) I believe this patch does apply to .Net 1.1.
2) No, it does not solve the other cases since they are

mostly a result of user configuration problems.

3) Yes, the fix does apply when the client is not an

ASP.NET client.

For the issues you are having, you should verify that

your proxy settings are correct and that you are
correctly specifying a valid server. Quite possibly
these two issues

are related. If you are still having these problems, I

might be able to help if you provide some more details.

Let me know if you have further questions.

Daniel Roth
--------------------
X-Tomcat-ID: 809948391
References: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----

=_NextPart_0001 _6ADD76EC"

Content-Transfer-Encoding: 7bit
From: da****@online.m icrosoft.com (Daniel Roth [MSFT])
Organization : Microsoft
Date: Thu, 26 Aug 2004 23:37:23 GMT
Subject: RE: Underlying connection was closed
X-Tomcat-NG:

microsoft.publi c.dotnet.framew ork.webservices

Message-ID: <np************ **@cpmsftngxa10 .phx.gbl>
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Lines: 120
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl

microsoft.publi c.dotnet.framew ork.webservices :6375

NNTP-Posting-Host: TOMCATIMPORT1 10.201.218.122

Hi Rajeev,

I am in the process of investigating your questions. I

will get back to you as soon as I have an answer.

Daniel Roth

--------------------
Content-Class: urn:content-classes:message
From: "Rajeev Tipnis" <rt*****@convoq .com>
Sender: "Rajeev Tipnis" <rt*****@convoq .com>
Subject: Underlying connection was closed
Date: Thu, 19 Aug 2004 14:59:51 -0700
Lines: 23
Message-ID: <99************ *************** *@phx.gbl>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Newsreader: Microsoft CDO for Windows 2000
Thread-Index: AcSGN9pVbhl3R/z4Q7iWwoav0NPNX g==
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
Newsgroups: microsoft.publi c.dotnet.framew ork.webservices
Path: cpmsftngxa10.ph x.gbl
Xref: cpmsftngxa10.ph x.gbl

microsoft.publi c.dotnet.framew ork.webservices :6290

NNTP-Posting-Host: tk2msftngxa14.p hx.gbl 10.40.1.166
X-Tomcat-NG:

microsoft.publi c.dotnet.framew ork.webservices

http://support.microsoft.com/default.aspx?scid=kb;en-
us;819450

Questions:
1) Is this patch (fix) applicable to the .NET 1.1
Framework as well? That is, if we have Framework 1.1 (On
Win2K3/IIS6.0), do we still need this patch?

2) Does it specifically solve only this error:
The underlying connection was closed. An unexpected

error

occured on send.
and not these (which is what we are seeing):

- The underlying connection was closed. Unable to

connect

to the remote server. OR
- The underlying connection was closed. The proxy name
could not be resolved, verify correct proxy

configuration.

3) Would this not happen when the webservices client is
not an ASP.NET client?

Thanks
Rajeev

--

This posting is provided "AS IS" with no warranties, and

confers no rights. Use of included script samples are
subject to the terms specified at

http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all

responses to this message are best directed to the
newsgroup/thread from which they originated.

--

This posting is provided "AS IS" with no warranties, and

confers no rights. Use of included script samples are
subject to the terms specified at

http://www.microsoft.com/info/cpyright.htm

Note: For the benefit of the community-at-large, all

responses to this message are best directed to the
newsgroup/thread from which they originated.

[Brian D.]

We keep getting the same intermittent problem in production. We have
turned off HTTP KEEP ALIVES in IIS (this seems "hacky" to me).

we have a fat client using HTTP Binary remoting over port 80 to IIS.
the problem seems to emerge when we get about 50 users on the system
at once. (we have 2 load balanced servers that have low CPU
utilization) ...

the web servers are running W2K with the latest .NET 1.1 SP.

What gives???? stack below.

brian
Server stack trace:
at System.Runtime. Remoting.Channe ls.Http.HttpCli entTransportSin k.ProcessRespon seException(Web Exception
webException, HttpWebResponse & response)
at System.Runtime. Remoting.Channe ls.Http.HttpCli entTransportSin k.ProcessMessag e(IMessage
msg, ITransportHeade rs requestHeaders, Stream requestStream,
ITransportHeade rs& responseHeaders , Stream& responseStream)
at System.Runtime. Remoting.Channe ls.BinaryClient FormatterSink.S yncProcessMessa ge(IMessage
msg)

Exception rethrown at [0]:
at System.Runtime. Remoting.Proxie s.RealProxy.Han dleReturnMessag e(IMessage
reqMsg, IMessage retMsg)
at System.Runtime. Remoting.Proxie s.RealProxy.Pri vateInvoke(Mess ageData&
msgData, Int32 type)