467,074 Members | 884 Online
Bytes | Developer Community
Ask Question

Home New Posts Topics Members FAQ

Post your question to a community of 467,074 developers. It's quick & easy.

Performance issue of ASP .NET

I'm currently building a website for one of the client. There has been
few errors here and there, but just recently the problem is getting worse.

Basically the symptoms is that when the user try to access the page, it
takes really long time to load. However, after up to 1 hour, the website
will run fine again as normal.

This issue has been there with the site. I usually just ask the system
admin to restart the IIS Service. However, the problem starts to occur
quite frequently recently (up to 3 times a day and during business hour).

Therefore, I'm doing a load testing using the same application on a
different box which can be accessed internally. I try to hit the page as
hard as I can with the tools (1000 user accessing the pages
simultaneously), but I still can't have the same problem occuring on the
testing box.

However, 1 thing that I found is that, on the testing box, when the
Requests Current counter (ASP .NET Counter) can go up as high as 1000,
while Requests Executing (ASP .NET Application Counter) only up to 1.
This is not the case that happen on the production box. I called the
system admin to have a look at the performance counter, and he said that
the Requests Current counter is around 250 with Requests Executing is
about the same number (250).

I try to search the net and no luck to find the cause of the problem. If
anyone here could help, I'd be very grateful, since the problem is very
urgent and need to be fixed ASAP.

Cheers.
Nov 19 '05 #1
  • viewed: 1729
Share:
17 Replies
Hi Sir,

Welcome to ASPNET newsgroup.
From your description, you're encountering some low perfomance issue on
your asp.net application in production environment, the application. The
symptom occurs up to 3 times a day recently ,yes?

Based on my experience, such low performance problem which turn normal
after certain period is possibly caused by the asp.net process recycle and
restarting which will cause the ASP.NET runtime being initializing (also
pages recompiling...)
Have you checked the server's application event log to see whether there
some ASP.NET worker process recycles log entries according to the times
that problem occur?

In addition, for general performance throubleshooting, we can check the
following things:

1. Whether CPU utilize is high when low performance occurs(such as
throughput lowdown..). If not high, there must have some long run task
which blocks threads....

2. We can also check whether the slow responding is cauesd by asp.net
processing or the response's output... sometimes when we the response
page's size being too large will cause slow response ,too...

BTW, for the different result you got in test environment, I think maybe
related to the testing tool you used which is not quite accurate for the
product envionment.

Thanks,

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)

--------------------
| Date: Thu, 10 Nov 2005 16:07:01 +1100
| From: 57R4N63R <un*****@unknown.com>
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Subject: Performance issue of ASP .NET
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| Message-ID: <ui**************@TK2MSFTNGP09.phx.gbl>
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| Lines: 1
| Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:356917
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| I'm currently building a website for one of the client. There has been
| few errors here and there, but just recently the problem is getting worse.
|
| Basically the symptoms is that when the user try to access the page, it
| takes really long time to load. However, after up to 1 hour, the website
| will run fine again as normal.
|
| This issue has been there with the site. I usually just ask the system
| admin to restart the IIS Service. However, the problem starts to occur
| quite frequently recently (up to 3 times a day and during business hour).
|
| Therefore, I'm doing a load testing using the same application on a
| different box which can be accessed internally. I try to hit the page as
| hard as I can with the tools (1000 user accessing the pages
| simultaneously), but I still can't have the same problem occuring on the
| testing box.
|
| However, 1 thing that I found is that, on the testing box, when the
| Requests Current counter (ASP .NET Counter) can go up as high as 1000,
| while Requests Executing (ASP .NET Application Counter) only up to 1.
| This is not the case that happen on the production box. I called the
| system admin to have a look at the performance counter, and he said that
| the Requests Current counter is around 250 with Requests Executing is
| about the same number (250).
|
| I try to search the net and no luck to find the cause of the problem. If
| anyone here could help, I'd be very grateful, since the problem is very
| urgent and need to be fixed ASAP.
|
| Cheers.
|

Nov 19 '05 #2
57R4N63R wrote:
I'm currently building a website for one of the client. There has been
few errors here and there, but just recently the problem is getting worse.

Basically the symptoms is that when the user try to access the page, it
takes really long time to load. However, after up to 1 hour, the website
will run fine again as normal.

This issue has been there with the site. I usually just ask the system
admin to restart the IIS Service. However, the problem starts to occur
quite frequently recently (up to 3 times a day and during business hour).

Therefore, I'm doing a load testing using the same application on a
different box which can be accessed internally. I try to hit the page as
hard as I can with the tools (1000 user accessing the pages
simultaneously), but I still can't have the same problem occuring on the
testing box.

However, 1 thing that I found is that, on the testing box, when the
Requests Current counter (ASP .NET Counter) can go up as high as 1000,
while Requests Executing (ASP .NET Application Counter) only up to 1.
This is not the case that happen on the production box. I called the
system admin to have a look at the performance counter, and he said that
the Requests Current counter is around 250 with Requests Executing is
about the same number (250).

I try to search the net and no luck to find the cause of the problem. If
anyone here could help, I'd be very grateful, since the problem is very
urgent and need to be fixed ASAP.

Cheers.


Can anyone help?
Nov 19 '05 #3
Hi Steven,

Thanks for the reply. Could you tell me how to check the Server's
Application Log? Is it on Event Viewer > Application?

Regarding your troubleshooting, I don't think it's both issue. The CPU
utilization for pretty low (about 10% is as high as we get) and when the
site works fine, it actually can display the pages pretty fast on
Internal LAN.

Budhi
Nov 19 '05 #4
Thanks for your response Budhi,

Yes, I means the Application Event log. You can view them through the
eventvwr.exe.
By default, when ASP.NET worker process recycle, it will add an entry in
the application event log.

Thanks,

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)

--------------------
| Date: Fri, 11 Nov 2005 14:57:02 +1100
| From: 57R4N63R <un*****@unknown.com>
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| Subject: Re: Performance issue of ASP .NET
| References: <ui**************@TK2MSFTNGP09.phx.gbl>
<ZN**************@TK2MSFTNGXA02.phx.gbl>
| In-Reply-To: <ZN**************@TK2MSFTNGXA02.phx.gbl>
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| NNTP-Posting-Host: ppp157-241.static.internode.on.net
| Message-ID: <43******@duster.adelaide.on.net>
| X-Trace: duster.adelaide.on.net 1131681422 150.101.157.241 (11 Nov 2005
14:27:02 +1050)
| Lines: 11
| Path:
TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!newsfee d00.sul.t-online.de!t-onli
ne.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!news-south.connect.com
au!duster.adelaide.on.net!not-for-mail
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:357269
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| Hi Steven,
|
| Thanks for the reply. Could you tell me how to check the Server's
| Application Log? Is it on Event Viewer > Application?
|
| Regarding your troubleshooting, I don't think it's both issue. The CPU
| utilization for pretty low (about 10% is as high as we get) and when the
| site works fine, it actually can display the pages pretty fast on
| Internal LAN.
|
| Budhi
|

Nov 19 '05 #5
Hi Steven,

Is there any reason why ASP.NET worker process would recycle and restart?

Budhi

Nov 19 '05 #6
Hi Budhi,

So did you find any recycle log entries on the server? As for the recycle,
there could have several causes, and generally the following twos are
common ones:

1. Worker process's Memory usage exceed the processModel (iis5) or
application pool(iis6) 's limitation...

2. Worker process encountering deadlock and become no-responding after
certain timeout period.

We can find such setting in the machine.config's <processModel> (iis5) or
in the IIS6's application pool settings. So if the symtom in your problem
is really caused by worker process recycling, I think you can start
throubleshooting from the above things....

Thanks,

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)

--------------------
| Date: Tue, 15 Nov 2005 11:21:23 +1100
| From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| Reply-To: bu***@immersive.com.au
| Organization: Immersive Pty Ltd
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Subject: Re: Performance issue of ASP .NET
| References: <ui**************@TK2MSFTNGP09.phx.gbl>
<ZN**************@TK2MSFTNGXA02.phx.gbl> <43******@duster.adelaide.on.net>
<8v*************@TK2MSFTNGXA02.phx.gbl>
| In-Reply-To: <8v*************@TK2MSFTNGXA02.phx.gbl>
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| Message-ID: <ua**************@tk2msftngp13.phx.gbl>
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| Lines: 1
| Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!tk2msft ngp13.phx.gbl
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:358115
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| Hi Steven,
|
| Is there any reason why ASP.NET worker process would recycle and restart?
|
| Budhi
|
|

Nov 19 '05 #7
Hi Steven,

According to the system administrator, there is no recycling log entries
during the event where the application stop responding. The recycling
time is set to 29 hours (default setup) if that matters. However there
are few deadlock error entries on the error log. The system admin also
said that he found few people got the same problem on the net but no one
answer the question. The cause is mostly because they are trying to move
the application from Windows Server 2000 to Windows Server 2003. There
description of the error message is:

ISAPI 'C:\WINDOWS\Microsoft.NET\Framework\v1.1.4322\aspn et_isapi.dll'
reported itself as unhealthy for the following reason: 'Deadlock detected'.

Budhi
Nov 20 '05 #8
Budhi Saputra Prasetya wrote:
Hi Steven,

According to the system administrator, there is no recycling log entries
during the event where the application stop responding. The recycling
time is set to 29 hours (default setup) if that matters. However there
are few deadlock error entries on the error log. The system admin also
said that he found few people got the same problem on the net but no one
answer the question. The cause is mostly because they are trying to move
the application from Windows Server 2000 to Windows Server 2003. There
description of the error message is:

ISAPI 'C:\WINDOWS\Microsoft.NET\Framework\v1.1.4322\aspn et_isapi.dll'
reported itself as unhealthy for the following reason: 'Deadlock detected'.

Budhi


Hi Steven,

There is one thing that I found by accident. When i try to compile the
result on the server performance counter, I found that the Request
Executing and Request Current counter is stuck at 4, while there is no
more request is actually submitted to the server. Could you help me with
the issue of why this could happen?

Budhi
Nov 20 '05 #9
Thanks for your response Budhi,

So from your two response, I got that there're some deadlock info detected
from teh Eventlog and you also mentioned that you found there has 4 current
request and executing requests stuck on the server, so possibly they're the
ones which has run into deadlock, do you think so? If so, we may have to
do some code analysis on our application. Is there any pages in your
application which could be the potential of the cause, e.g, using some
resources or shared objects ?

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)

--------------------
| Date: Wed, 16 Nov 2005 16:35:48 +1100
| From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| Reply-To: bu***@immersive.com.au
| Organization: Immersive Pty Ltd
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Subject: Re: Performance issue of ASP .NET
| References: <ui**************@TK2MSFTNGP09.phx.gbl>
<ZN**************@TK2MSFTNGXA02.phx.gbl> <43******@duster.adelaide.on.net>
<8v*************@TK2MSFTNGXA02.phx.gbl>
<ua**************@tk2msftngp13.phx.gbl>
<gh**************@TK2MSFTNGXA02.phx.gbl>
<#8**************@tk2msftngp13.phx.gbl>
| In-Reply-To: <#8**************@tk2msftngp13.phx.gbl>
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| Message-ID: <#p**************@TK2MSFTNGP09.phx.gbl>
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| Lines: 1
| Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:358476
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| Budhi Saputra Prasetya wrote:
| > Hi Steven,
| >
| > According to the system administrator, there is no recycling log
entries
| > during the event where the application stop responding. The recycling
| > time is set to 29 hours (default setup) if that matters. However there
| > are few deadlock error entries on the error log. The system admin also
| > said that he found few people got the same problem on the net but no
one
| > answer the question. The cause is mostly because they are trying to
move
| > the application from Windows Server 2000 to Windows Server 2003. There
| > description of the error message is:
| >
| > ISAPI 'C:\WINDOWS\Microsoft.NET\Framework\v1.1.4322\aspn et_isapi.dll'
| > reported itself as unhealthy for the following reason: 'Deadlock
detected'.
| >
| > Budhi
|
| Hi Steven,
|
| There is one thing that I found by accident. When i try to compile the
| result on the server performance counter, I found that the Request
| Executing and Request Current counter is stuck at 4, while there is no
| more request is actually submitted to the server. Could you help me with
| the issue of why this could happen?
|
| Budhi
|

Nov 20 '05 #10
Steven Cheng[MSFT] wrote:
Thanks for your response Budhi,

So from your two response, I got that there're some deadlock info detected
from teh Eventlog and you also mentioned that you found there has 4 current
request and executing requests stuck on the server, so possibly they're the
ones which has run into deadlock, do you think so? If so, we may have to
do some code analysis on our application. Is there any pages in your
application which could be the potential of the cause, e.g, using some
resources or shared objects ?

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)

--------------------
| Date: Wed, 16 Nov 2005 16:35:48 +1100
| From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| Reply-To: bu***@immersive.com.au
| Organization: Immersive Pty Ltd
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Subject: Re: Performance issue of ASP .NET
| References: <ui**************@TK2MSFTNGP09.phx.gbl>
<ZN**************@TK2MSFTNGXA02.phx.gbl> <43******@duster.adelaide.on.net>
<8v*************@TK2MSFTNGXA02.phx.gbl>
<ua**************@tk2msftngp13.phx.gbl>
<gh**************@TK2MSFTNGXA02.phx.gbl>
<#8**************@tk2msftngp13.phx.gbl>
| In-Reply-To: <#8**************@tk2msftngp13.phx.gbl>
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| Message-ID: <#p**************@TK2MSFTNGP09.phx.gbl>
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| Lines: 1
| Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:358476
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| Budhi Saputra Prasetya wrote:
| > Hi Steven,
| >
| > According to the system administrator, there is no recycling log
entries
| > during the event where the application stop responding. The recycling
| > time is set to 29 hours (default setup) if that matters. However there
| > are few deadlock error entries on the error log. The system admin also
| > said that he found few people got the same problem on the net but no
one
| > answer the question. The cause is mostly because they are trying to
move
| > the application from Windows Server 2000 to Windows Server 2003. There
| > description of the error message is:
| >
| > ISAPI 'C:\WINDOWS\Microsoft.NET\Framework\v1.1.4322\aspn et_isapi.dll'
| > reported itself as unhealthy for the following reason: 'Deadlock
detected'.
| >
| > Budhi
|
| Hi Steven,
|
| There is one thing that I found by accident. When i try to compile the
| result on the server performance counter, I found that the Request
| Executing and Request Current counter is stuck at 4, while there is no
| more request is actually submitted to the server. Could you help me with
| the issue of why this could happen?
|
| Budhi
|


Hi Steven,

There is only 1 shared objects (Application Object), however the locking
only happen when I want to check whether a new generated id is a
duplicate id, because the id is shared among all tables. As soon as the
id is generated I do unlock.

Basically the pseudocode looks like:

1. Generate id
2. Lock the Application Object
3. Does it exist in the list, if not put a new id into the list
4. Unlock the Application Object
5. If duplicate go back to step 1 else continue with next step

As you see it's very short period amount of time taken for lock and
unlock. But do you know why the counter can be stuck if that's the only
resources I lock, cos I don't see how that lock and unlock in this case
can cause a problem since I'm doing it only on 1 resource. I did the
stress testing few times already and there wasn't any problem before.
When the stress testing finished, the counter always go back to 0.

Budhi
Nov 20 '05 #11
Thanks for your response Budhi,

I'm also not very sure about the perf counter behavior you mentioned. Also,
the appication object locking is just a potential cause. It'll be better if
we could isolate the problem to some certain page.

Regards,

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)
--------------------
| Date: Thu, 17 Nov 2005 11:58:20 +1100
| From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| Reply-To: bu***@immersive.com.au
| Organization: Immersive Pty Ltd
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Subject: Re: Performance issue of ASP .NET
| References: <ui**************@TK2MSFTNGP09.phx.gbl>
<ZN**************@TK2MSFTNGXA02.phx.gbl> <43******@duster.adelaide.on.net>
<8v*************@TK2MSFTNGXA02.phx.gbl>
<ua**************@tk2msftngp13.phx.gbl>
<gh**************@TK2MSFTNGXA02.phx.gbl>
<#8**************@tk2msftngp13.phx.gbl>
<#p**************@TK2MSFTNGP09.phx.gbl>
<pB**************@TK2MSFTNGXA02.phx.gbl>
| In-Reply-To: <pB**************@TK2MSFTNGXA02.phx.gbl>
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| Message-ID: <Oe*************@TK2MSFTNGP09.phx.gbl>
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| Lines: 1
| Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:358793
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| Steven Cheng[MSFT] wrote:
| > Thanks for your response Budhi,
| >
| > So from your two response, I got that there're some deadlock info
detected
| > from teh Eventlog and you also mentioned that you found there has 4
current
| > request and executing requests stuck on the server, so possibly they're
the
| > ones which has run into deadlock, do you think so? If so, we may have
to
| > do some code analysis on our application. Is there any pages in your
| > application which could be the potential of the cause, e.g, using some
| > resources or shared objects ?
| >
| > Steven Cheng
| > Microsoft Online Support
| >
| > Get Secure! www.microsoft.com/security
| > (This posting is provided "AS IS", with no warranties, and confers no
| > rights.)
| >
| >
| >
| > --------------------
| > | Date: Wed, 16 Nov 2005 16:35:48 +1100
| > | From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| > | Reply-To: bu***@immersive.com.au
| > | Organization: Immersive Pty Ltd
| > | User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| > | X-Accept-Language: en-us, en
| > | MIME-Version: 1.0
| > | Subject: Re: Performance issue of ASP .NET
| > | References: <ui**************@TK2MSFTNGP09.phx.gbl>
| > <ZN**************@TK2MSFTNGXA02.phx.gbl>
<43******@duster.adelaide.on.net>
| > <8v*************@TK2MSFTNGXA02.phx.gbl>
| > <ua**************@tk2msftngp13.phx.gbl>
| > <gh**************@TK2MSFTNGXA02.phx.gbl>
| > <#8**************@tk2msftngp13.phx.gbl>
| > | In-Reply-To: <#8**************@tk2msftngp13.phx.gbl>
| > | Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| > | Content-Transfer-Encoding: 7bit
| > | Message-ID: <#p**************@TK2MSFTNGP09.phx.gbl>
| > | Newsgroups: microsoft.public.dotnet.framework.aspnet
| > | NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| > | Lines: 1
| > | Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| > | Xref: TK2MSFTNGXA02.phx.gbl
| > microsoft.public.dotnet.framework.aspnet:358476
| > | X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
| > |
| > | Budhi Saputra Prasetya wrote:
| > | > Hi Steven,
| > | >
| > | > According to the system administrator, there is no recycling log
| > entries
| > | > during the event where the application stop responding. The
recycling
| > | > time is set to 29 hours (default setup) if that matters. However
there
| > | > are few deadlock error entries on the error log. The system admin
also
| > | > said that he found few people got the same problem on the net but
no
| > one
| > | > answer the question. The cause is mostly because they are trying to
| > move
| > | > the application from Windows Server 2000 to Windows Server 2003.
There
| > | > description of the error message is:
| > | >
| > | > ISAPI
'C:\WINDOWS\Microsoft.NET\Framework\v1.1.4322\aspn et_isapi.dll'
| > | > reported itself as unhealthy for the following reason: 'Deadlock
| > detected'.
| > | >
| > | > Budhi
| > |
| > | Hi Steven,
| > |
| > | There is one thing that I found by accident. When i try to compile
the
| > | result on the server performance counter, I found that the Request
| > | Executing and Request Current counter is stuck at 4, while there is
no
| > | more request is actually submitted to the server. Could you help me
with
| > | the issue of why this could happen?
| > |
| > | Budhi
| > |
| >
|
| Hi Steven,
|
| There is only 1 shared objects (Application Object), however the locking
| only happen when I want to check whether a new generated id is a
| duplicate id, because the id is shared among all tables. As soon as the
| id is generated I do unlock.
|
| Basically the pseudocode looks like:
|
| 1. Generate id
| 2. Lock the Application Object
| 3. Does it exist in the list, if not put a new id into the list
| 4. Unlock the Application Object
| 5. If duplicate go back to step 1 else continue with next step
|
| As you see it's very short period amount of time taken for lock and
| unlock. But do you know why the counter can be stuck if that's the only
| resources I lock, cos I don't see how that lock and unlock in this case
| can cause a problem since I'm doing it only on 1 resource. I did the
| stress testing few times already and there wasn't any problem before.
| When the stress testing finished, the counter always go back to 0.
|
| Budhi
|

Nov 20 '05 #12
Budhi,

If you lock a resource, do some steps, get an exception while the resource is locked,
and then exit to an exception handler you will not unlock the resource.

Is it possible you are returning to a higher level due to an exception or some other
reason while you have a resource lock in effect?

Budhi Saputra Prasetya wrote:

Hi Steven,

There is only 1 shared objects (Application Object), however the locking
only happen when I want to check whether a new generated id is a
duplicate id, because the id is shared among all tables. As soon as the
id is generated I do unlock.

Basically the pseudocode looks like:

1. Generate id
2. Lock the Application Object
3. Does it exist in the list, if not put a new id into the list
4. Unlock the Application Object
5. If duplicate go back to step 1 else continue with next step

As you see it's very short period amount of time taken for lock and
unlock. But do you know why the counter can be stuck if that's the only
resources I lock, cos I don't see how that lock and unlock in this case
can cause a problem since I'm doing it only on 1 resource. I did the
stress testing few times already and there wasn't any problem before.
When the stress testing finished, the counter always go back to 0.

Budhi

Nov 20 '05 #13
Randall Parker wrote:
Budhi,

If you lock a resource, do some steps, get an exception while the
resource is locked, and then exit to an exception handler you will not
unlock the resource.

Is it possible you are returning to a higher level due to an exception
or some other reason while you have a resource lock in effect?

Budhi Saputra Prasetya wrote:

Hi Steven,

There is only 1 shared objects (Application Object), however the
locking only happen when I want to check whether a new generated id is
a duplicate id, because the id is shared among all tables. As soon as
the id is generated I do unlock.

Basically the pseudocode looks like:

1. Generate id
2. Lock the Application Object
3. Does it exist in the list, if not put a new id into the list
4. Unlock the Application Object
5. If duplicate go back to step 1 else continue with next step

As you see it's very short period amount of time taken for lock and
unlock. But do you know why the counter can be stuck if that's the
only resources I lock, cos I don't see how that lock and unlock in
this case can cause a problem since I'm doing it only on 1 resource. I
did the stress testing few times already and there wasn't any problem
before. When the stress testing finished, the counter always go back
to 0.

Budhi


Hi Randall,

I understand what you were saying. However, I'm pretty sure I handled
the exception during the Application Object locking. This is the code
that I have to check if the object id already exist within the collection:

Public Function IsObjectIdExist(ByVal ObjectId As String) As Boolean
Try
Return (Not (_ObjectIdCollection.Item(ObjectId) Is Nothing))
Catch ex As ArgumentException
Return False
Catch ex As IndexOutOfRangeException
Return False
End Try
End Function

I still can't see the problem through the Application Object locking.
I'm suspecting it might be a database locking issue, because I don't
have any other locking procedure within application level.

Budhi
Nov 23 '05 #14
Steven Cheng[MSFT] wrote:
Thanks for your response Budhi,

I'm also not very sure about the perf counter behavior you mentioned. Also,
the appication object locking is just a potential cause. It'll be better if
we could isolate the problem to some certain page.

Regards,

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)
--------------------
| Date: Thu, 17 Nov 2005 11:58:20 +1100
| From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| Reply-To: bu***@immersive.com.au
| Organization: Immersive Pty Ltd
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Subject: Re: Performance issue of ASP .NET
| References: <ui**************@TK2MSFTNGP09.phx.gbl>
<ZN**************@TK2MSFTNGXA02.phx.gbl> <43******@duster.adelaide.on.net>
<8v*************@TK2MSFTNGXA02.phx.gbl>
<ua**************@tk2msftngp13.phx.gbl>
<gh**************@TK2MSFTNGXA02.phx.gbl>
<#8**************@tk2msftngp13.phx.gbl>
<#p**************@TK2MSFTNGP09.phx.gbl>
<pB**************@TK2MSFTNGXA02.phx.gbl>
| In-Reply-To: <pB**************@TK2MSFTNGXA02.phx.gbl>
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| Message-ID: <Oe*************@TK2MSFTNGP09.phx.gbl>
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| Lines: 1
| Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:358793
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| Steven Cheng[MSFT] wrote:
| > Thanks for your response Budhi,
| >
| > So from your two response, I got that there're some deadlock info
detected
| > from teh Eventlog and you also mentioned that you found there has 4
current
| > request and executing requests stuck on the server, so possibly they're
the
| > ones which has run into deadlock, do you think so? If so, we may have
to
| > do some code analysis on our application. Is there any pages in your
| > application which could be the potential of the cause, e.g, using some
| > resources or shared objects ?
| >
| > Steven Cheng
| > Microsoft Online Support
| >
| > Get Secure! www.microsoft.com/security
| > (This posting is provided "AS IS", with no warranties, and confers no
| > rights.)
| >
| >
| >
| > --------------------
| > | Date: Wed, 16 Nov 2005 16:35:48 +1100
| > | From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| > | Reply-To: bu***@immersive.com.au
| > | Organization: Immersive Pty Ltd
| > | User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| > | X-Accept-Language: en-us, en
| > | MIME-Version: 1.0
| > | Subject: Re: Performance issue of ASP .NET
| > | References: <ui**************@TK2MSFTNGP09.phx.gbl>
| > <ZN**************@TK2MSFTNGXA02.phx.gbl>
<43******@duster.adelaide.on.net>
| > <8v*************@TK2MSFTNGXA02.phx.gbl>
| > <ua**************@tk2msftngp13.phx.gbl>
| > <gh**************@TK2MSFTNGXA02.phx.gbl>
| > <#8**************@tk2msftngp13.phx.gbl>
| > | In-Reply-To: <#8**************@tk2msftngp13.phx.gbl>
| > | Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| > | Content-Transfer-Encoding: 7bit
| > | Message-ID: <#p**************@TK2MSFTNGP09.phx.gbl>
| > | Newsgroups: microsoft.public.dotnet.framework.aspnet
| > | NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| > | Lines: 1
| > | Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| > | Xref: TK2MSFTNGXA02.phx.gbl
| > microsoft.public.dotnet.framework.aspnet:358476
| > | X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
| > |
| > | Budhi Saputra Prasetya wrote:
| > | > Hi Steven,
| > | >
| > | > According to the system administrator, there is no recycling log
| > entries
| > | > during the event where the application stop responding. The
recycling
| > | > time is set to 29 hours (default setup) if that matters. However
there
| > | > are few deadlock error entries on the error log. The system admin
also
| > | > said that he found few people got the same problem on the net but
no
| > one
| > | > answer the question. The cause is mostly because they are trying to
| > move
| > | > the application from Windows Server 2000 to Windows Server 2003.
There
| > | > description of the error message is:
| > | >
| > | > ISAPI
'C:\WINDOWS\Microsoft.NET\Framework\v1.1.4322\aspn et_isapi.dll'
| > | > reported itself as unhealthy for the following reason: 'Deadlock
| > detected'.
| > | >
| > | > Budhi
| > |
| > | Hi Steven,
| > |
| > | There is one thing that I found by accident. When i try to compile
the
| > | result on the server performance counter, I found that the Request
| > | Executing and Request Current counter is stuck at 4, while there is
no
| > | more request is actually submitted to the server. Could you help me
with
| > | the issue of why this could happen?
| > |
| > | Budhi
| > |
| >
|
| Hi Steven,
|
| There is only 1 shared objects (Application Object), however the locking
| only happen when I want to check whether a new generated id is a
| duplicate id, because the id is shared among all tables. As soon as the
| id is generated I do unlock.
|
| Basically the pseudocode looks like:
|
| 1. Generate id
| 2. Lock the Application Object
| 3. Does it exist in the list, if not put a new id into the list
| 4. Unlock the Application Object
| 5. If duplicate go back to step 1 else continue with next step
|
| As you see it's very short period amount of time taken for lock and
| unlock. But do you know why the counter can be stuck if that's the only
| resources I lock, cos I don't see how that lock and unlock in this case
| can cause a problem since I'm doing it only on 1 resource. I did the
| stress testing few times already and there wasn't any problem before.
| When the stress testing finished, the counter always go back to 0.
|
| Budhi
|


Hi Steven,

Sorry just came back from 2 days break from work. :) About the problem,
I started to think that it might be because of the database locking
procedure. Could it be possible since I don't do any specific setup on
database level. By the way, we are using MS SQL Server 2000 if that matter.

Budhi
Nov 23 '05 #15
Thanks for your followup Budhi,

I think SqlServer version should not matter here. Make sure the database
server and sqlserver has installed the latest patch... Also, is the
database being called frequently in your application?

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)
--------------------
| Date: Wed, 23 Nov 2005 10:25:06 +1100
| From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| Reply-To: bu***@immersive.com.au
| Organization: Immersive Pty Ltd
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Subject: Re: Performance issue of ASP .NET
| References: <ui**************@TK2MSFTNGP09.phx.gbl>
<ZN**************@TK2MSFTNGXA02.phx.gbl> <43******@duster.adelaide.on.net>
<8v*************@TK2MSFTNGXA02.phx.gbl>
<ua**************@tk2msftngp13.phx.gbl>
<gh**************@TK2MSFTNGXA02.phx.gbl>
<#8**************@tk2msftngp13.phx.gbl>
<#p**************@TK2MSFTNGP09.phx.gbl>
<pB**************@TK2MSFTNGXA02.phx.gbl>
<Oe*************@TK2MSFTNGP09.phx.gbl>
<pg**************@TK2MSFTNGXA02.phx.gbl>
| In-Reply-To: <pg**************@TK2MSFTNGXA02.phx.gbl>
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| Message-ID: <#C**************@TK2MSFTNGP14.phx.gbl>
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| Lines: 1
| Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP14.phx.gbl
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:360284
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| Steven Cheng[MSFT] wrote:
| > Thanks for your response Budhi,
| >
| > I'm also not very sure about the perf counter behavior you mentioned.
Also,
| > the appication object locking is just a potential cause. It'll be
better if
| > we could isolate the problem to some certain page.
| >
| > Regards,
| >
| > Steven Cheng
| > Microsoft Online Support
| >
| > Get Secure! www.microsoft.com/security
| > (This posting is provided "AS IS", with no warranties, and confers no
| > rights.)
| >
| >
| > --------------------
| > | Date: Thu, 17 Nov 2005 11:58:20 +1100
| > | From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| > | Reply-To: bu***@immersive.com.au
| > | Organization: Immersive Pty Ltd
| > | User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| > | X-Accept-Language: en-us, en
| > | MIME-Version: 1.0
| > | Subject: Re: Performance issue of ASP .NET
| > | References: <ui**************@TK2MSFTNGP09.phx.gbl>
| > <ZN**************@TK2MSFTNGXA02.phx.gbl>
<43******@duster.adelaide.on.net>
| > <8v*************@TK2MSFTNGXA02.phx.gbl>
| > <ua**************@tk2msftngp13.phx.gbl>
| > <gh**************@TK2MSFTNGXA02.phx.gbl>
| > <#8**************@tk2msftngp13.phx.gbl>
| > <#p**************@TK2MSFTNGP09.phx.gbl>
| > <pB**************@TK2MSFTNGXA02.phx.gbl>
| > | In-Reply-To: <pB**************@TK2MSFTNGXA02.phx.gbl>
| > | Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| > | Content-Transfer-Encoding: 7bit
| > | Message-ID: <Oe*************@TK2MSFTNGP09.phx.gbl>
| > | Newsgroups: microsoft.public.dotnet.framework.aspnet
| > | NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| > | Lines: 1
| > | Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| > | Xref: TK2MSFTNGXA02.phx.gbl
| > microsoft.public.dotnet.framework.aspnet:358793
| > | X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
| > |
| > | Steven Cheng[MSFT] wrote:
| > | > Thanks for your response Budhi,
| > | >
| > | > So from your two response, I got that there're some deadlock info
| > detected
| > | > from teh Eventlog and you also mentioned that you found there has 4
| > current
| > | > request and executing requests stuck on the server, so possibly
they're
| > the
| > | > ones which has run into deadlock, do you think so? If so, we may
have
| > to
| > | > do some code analysis on our application. Is there any pages in
your
| > | > application which could be the potential of the cause, e.g, using
some
| > | > resources or shared objects ?
| > | >
| > | > Steven Cheng
| > | > Microsoft Online Support
| > | >
| > | > Get Secure! www.microsoft.com/security
| > | > (This posting is provided "AS IS", with no warranties, and confers
no
| > | > rights.)
| > | >
| > | >
| > | >
| > | > --------------------
| > | > | Date: Wed, 16 Nov 2005 16:35:48 +1100
| > | > | From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| > | > | Reply-To: bu***@immersive.com.au
| > | > | Organization: Immersive Pty Ltd
| > | > | User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| > | > | X-Accept-Language: en-us, en
| > | > | MIME-Version: 1.0
| > | > | Subject: Re: Performance issue of ASP .NET
| > | > | References: <ui**************@TK2MSFTNGP09.phx.gbl>
| > | > <ZN**************@TK2MSFTNGXA02.phx.gbl>
| > <43******@duster.adelaide.on.net>
| > | > <8v*************@TK2MSFTNGXA02.phx.gbl>
| > | > <ua**************@tk2msftngp13.phx.gbl>
| > | > <gh**************@TK2MSFTNGXA02.phx.gbl>
| > | > <#8**************@tk2msftngp13.phx.gbl>
| > | > | In-Reply-To: <#8**************@tk2msftngp13.phx.gbl>
| > | > | Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| > | > | Content-Transfer-Encoding: 7bit
| > | > | Message-ID: <#p**************@TK2MSFTNGP09.phx.gbl>
| > | > | Newsgroups: microsoft.public.dotnet.framework.aspnet
| > | > | NNTP-Posting-Host: ppp157-241.static.internode.on.net
150.101.157.241
| > | > | Lines: 1
| > | > | Path:
TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| > | > | Xref: TK2MSFTNGXA02.phx.gbl
| > | > microsoft.public.dotnet.framework.aspnet:358476
| > | > | X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
| > | > |
| > | > | Budhi Saputra Prasetya wrote:
| > | > | > Hi Steven,
| > | > | >
| > | > | > According to the system administrator, there is no recycling
log
| > | > entries
| > | > | > during the event where the application stop responding. The
| > recycling
| > | > | > time is set to 29 hours (default setup) if that matters.
However
| > there
| > | > | > are few deadlock error entries on the error log. The system
admin
| > also
| > | > | > said that he found few people got the same problem on the net
but
| > no
| > | > one
| > | > | > answer the question. The cause is mostly because they are
trying to
| > | > move
| > | > | > the application from Windows Server 2000 to Windows Server
2003.
| > There
| > | > | > description of the error message is:
| > | > | >
| > | > | > ISAPI
| > 'C:\WINDOWS\Microsoft.NET\Framework\v1.1.4322\aspn et_isapi.dll'
| > | > | > reported itself as unhealthy for the following reason:
'Deadlock
| > | > detected'.
| > | > | >
| > | > | > Budhi
| > | > |
| > | > | Hi Steven,
| > | > |
| > | > | There is one thing that I found by accident. When i try to
compile
| > the
| > | > | result on the server performance counter, I found that the
Request
| > | > | Executing and Request Current counter is stuck at 4, while there
is
| > no
| > | > | more request is actually submitted to the server. Could you help
me
| > with
| > | > | the issue of why this could happen?
| > | > |
| > | > | Budhi
| > | > |
| > | >
| > |
| > | Hi Steven,
| > |
| > | There is only 1 shared objects (Application Object), however the
locking
| > | only happen when I want to check whether a new generated id is a
| > | duplicate id, because the id is shared among all tables. As soon as
the
| > | id is generated I do unlock.
| > |
| > | Basically the pseudocode looks like:
| > |
| > | 1. Generate id
| > | 2. Lock the Application Object
| > | 3. Does it exist in the list, if not put a new id into the list
| > | 4. Unlock the Application Object
| > | 5. If duplicate go back to step 1 else continue with next step
| > |
| > | As you see it's very short period amount of time taken for lock and
| > | unlock. But do you know why the counter can be stuck if that's the
only
| > | resources I lock, cos I don't see how that lock and unlock in this
case
| > | can cause a problem since I'm doing it only on 1 resource. I did the
| > | stress testing few times already and there wasn't any problem before.
| > | When the stress testing finished, the counter always go back to 0.
| > |
| > | Budhi
| > |
| >
|
| Hi Steven,
|
| Sorry just came back from 2 days break from work. :) About the problem,
| I started to think that it might be because of the database locking
| procedure. Could it be possible since I don't do any specific setup on
| database level. By the way, we are using MS SQL Server 2000 if that
matter.
|
| Budhi
|

Nov 23 '05 #16
Hi Steven,

The database is called a lot of times. To request 1 page, the software
will call the database about 5-10 times to retrieve the data.
Personally, I don't like the design, but I wasn't the designer. And also
they aren't grouped together as a transaction. So we have individual SQL
Query trying to access the database. Will this cause the problem?

I've been trying to do another stress testing, but I still have no luck
in reproducing request executing counter locked. When the testing
finished, the counter still goes back to 0 and there is no log in the
Application Log saying that ASP .NET recycling because of deadlock.

Budhi

Steven Cheng[MSFT] wrote:
Thanks for your followup Budhi,

I think SqlServer version should not matter here. Make sure the database
server and sqlserver has installed the latest patch... Also, is the
database being called frequently in your application?

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)
--------------------
| Date: Wed, 23 Nov 2005 10:25:06 +1100
| From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| Reply-To: bu***@immersive.com.au
| Organization: Immersive Pty Ltd
| User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| X-Accept-Language: en-us, en
| MIME-Version: 1.0
| Subject: Re: Performance issue of ASP .NET
| References: <ui**************@TK2MSFTNGP09.phx.gbl>
<ZN**************@TK2MSFTNGXA02.phx.gbl> <43******@duster.adelaide.on.net>
<8v*************@TK2MSFTNGXA02.phx.gbl>
<ua**************@tk2msftngp13.phx.gbl>
<gh**************@TK2MSFTNGXA02.phx.gbl>
<#8**************@tk2msftngp13.phx.gbl>
<#p**************@TK2MSFTNGP09.phx.gbl>
<pB**************@TK2MSFTNGXA02.phx.gbl>
<Oe*************@TK2MSFTNGP09.phx.gbl>
<pg**************@TK2MSFTNGXA02.phx.gbl>
| In-Reply-To: <pg**************@TK2MSFTNGXA02.phx.gbl>
| Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| Content-Transfer-Encoding: 7bit
| Message-ID: <#C**************@TK2MSFTNGP14.phx.gbl>
| Newsgroups: microsoft.public.dotnet.framework.aspnet
| NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| Lines: 1
| Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP14.phx.gbl
| Xref: TK2MSFTNGXA02.phx.gbl
microsoft.public.dotnet.framework.aspnet:360284
| X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
|
| Steven Cheng[MSFT] wrote:
| > Thanks for your response Budhi,
| >
| > I'm also not very sure about the perf counter behavior you mentioned.
Also,
| > the appication object locking is just a potential cause. It'll be
better if
| > we could isolate the problem to some certain page.
| >
| > Regards,
| >
| > Steven Cheng
| > Microsoft Online Support
| >
| > Get Secure! www.microsoft.com/security
| > (This posting is provided "AS IS", with no warranties, and confers no
| > rights.)
| >
| >
| > --------------------
| > | Date: Thu, 17 Nov 2005 11:58:20 +1100
| > | From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| > | Reply-To: bu***@immersive.com.au
| > | Organization: Immersive Pty Ltd
| > | User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| > | X-Accept-Language: en-us, en
| > | MIME-Version: 1.0
| > | Subject: Re: Performance issue of ASP .NET
| > | References: <ui**************@TK2MSFTNGP09.phx.gbl>
| > <ZN**************@TK2MSFTNGXA02.phx.gbl>
<43******@duster.adelaide.on.net>
| > <8v*************@TK2MSFTNGXA02.phx.gbl>
| > <ua**************@tk2msftngp13.phx.gbl>
| > <gh**************@TK2MSFTNGXA02.phx.gbl>
| > <#8**************@tk2msftngp13.phx.gbl>
| > <#p**************@TK2MSFTNGP09.phx.gbl>
| > <pB**************@TK2MSFTNGXA02.phx.gbl>
| > | In-Reply-To: <pB**************@TK2MSFTNGXA02.phx.gbl>
| > | Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| > | Content-Transfer-Encoding: 7bit
| > | Message-ID: <Oe*************@TK2MSFTNGP09.phx.gbl>
| > | Newsgroups: microsoft.public.dotnet.framework.aspnet
| > | NNTP-Posting-Host: ppp157-241.static.internode.on.net 150.101.157.241
| > | Lines: 1
| > | Path: TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| > | Xref: TK2MSFTNGXA02.phx.gbl
| > microsoft.public.dotnet.framework.aspnet:358793
| > | X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
| > |
| > | Steven Cheng[MSFT] wrote:
| > | > Thanks for your response Budhi,
| > | >
| > | > So from your two response, I got that there're some deadlock info
| > detected
| > | > from teh Eventlog and you also mentioned that you found there has 4
| > current
| > | > request and executing requests stuck on the server, so possibly
they're
| > the
| > | > ones which has run into deadlock, do you think so? If so, we may
have
| > to
| > | > do some code analysis on our application. Is there any pages in
your
| > | > application which could be the potential of the cause, e.g, using
some
| > | > resources or shared objects ?
| > | >
| > | > Steven Cheng
| > | > Microsoft Online Support
| > | >
| > | > Get Secure! www.microsoft.com/security
| > | > (This posting is provided "AS IS", with no warranties, and confers
no
| > | > rights.)
| > | >
| > | >
| > | >
| > | > --------------------
| > | > | Date: Wed, 16 Nov 2005 16:35:48 +1100
| > | > | From: Budhi Saputra Prasetya <bu***@immersive.com.au>
| > | > | Reply-To: bu***@immersive.com.au
| > | > | Organization: Immersive Pty Ltd
| > | > | User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
| > | > | X-Accept-Language: en-us, en
| > | > | MIME-Version: 1.0
| > | > | Subject: Re: Performance issue of ASP .NET
| > | > | References: <ui**************@TK2MSFTNGP09.phx.gbl>
| > | > <ZN**************@TK2MSFTNGXA02.phx.gbl>
| > <43******@duster.adelaide.on.net>
| > | > <8v*************@TK2MSFTNGXA02.phx.gbl>
| > | > <ua**************@tk2msftngp13.phx.gbl>
| > | > <gh**************@TK2MSFTNGXA02.phx.gbl>
| > | > <#8**************@tk2msftngp13.phx.gbl>
| > | > | In-Reply-To: <#8**************@tk2msftngp13.phx.gbl>
| > | > | Content-Type: text/plain; charset=ISO-8859-1; format=flowed
| > | > | Content-Transfer-Encoding: 7bit
| > | > | Message-ID: <#p**************@TK2MSFTNGP09.phx.gbl>
| > | > | Newsgroups: microsoft.public.dotnet.framework.aspnet
| > | > | NNTP-Posting-Host: ppp157-241.static.internode.on.net
150.101.157.241
| > | > | Lines: 1
| > | > | Path:
TK2MSFTNGXA02.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFT NGP09.phx.gbl
| > | > | Xref: TK2MSFTNGXA02.phx.gbl
| > | > microsoft.public.dotnet.framework.aspnet:358476
| > | > | X-Tomcat-NG: microsoft.public.dotnet.framework.aspnet
| > | > |
| > | > | Budhi Saputra Prasetya wrote:
| > | > | > Hi Steven,
| > | > | >
| > | > | > According to the system administrator, there is no recycling
log
| > | > entries
| > | > | > during the event where the application stop responding. The
| > recycling
| > | > | > time is set to 29 hours (default setup) if that matters.
However
| > there
| > | > | > are few deadlock error entries on the error log. The system
admin
| > also
| > | > | > said that he found few people got the same problem on the net
but
| > no
| > | > one
| > | > | > answer the question. The cause is mostly because they are
trying to
| > | > move
| > | > | > the application from Windows Server 2000 to Windows Server
2003.
| > There
| > | > | > description of the error message is:
| > | > | >
| > | > | > ISAPI
| > 'C:\WINDOWS\Microsoft.NET\Framework\v1.1.4322\aspn et_isapi.dll'
| > | > | > reported itself as unhealthy for the following reason:
'Deadlock
| > | > detected'.
| > | > | >
| > | > | > Budhi
| > | > |
| > | > | Hi Steven,
| > | > |
| > | > | There is one thing that I found by accident. When i try to
compile
| > the
| > | > | result on the server performance counter, I found that the
Request
| > | > | Executing and Request Current counter is stuck at 4, while there
is
| > no
| > | > | more request is actually submitted to the server. Could you help
me
| > with
| > | > | the issue of why this could happen?
| > | > |
| > | > | Budhi
| > | > |
| > | >
| > |
| > | Hi Steven,
| > |
| > | There is only 1 shared objects (Application Object), however the
locking
| > | only happen when I want to check whether a new generated id is a
| > | duplicate id, because the id is shared among all tables. As soon as
the
| > | id is generated I do unlock.
| > |
| > | Basically the pseudocode looks like:
| > |
| > | 1. Generate id
| > | 2. Lock the Application Object
| > | 3. Does it exist in the list, if not put a new id into the list
| > | 4. Unlock the Application Object
| > | 5. If duplicate go back to step 1 else continue with next step
| > |
| > | As you see it's very short period amount of time taken for lock and
| > | unlock. But do you know why the counter can be stuck if that's the
only
| > | resources I lock, cos I don't see how that lock and unlock in this
case
| > | can cause a problem since I'm doing it only on 1 resource. I did the
| > | stress testing few times already and there wasn't any problem before.
| > | When the stress testing finished, the counter always go back to 0.
| > |
| > | Budhi
| > |
| >
|
| Hi Steven,
|
| Sorry just came back from 2 days break from work. :) About the problem,
| I started to think that it might be because of the database locking
| procedure. Could it be possible since I don't do any specific setup on
| database level. By the way, we are using MS SQL Server 2000 if that
matter.
|
| Budhi
|

Nov 24 '05 #17
Thanks for your reply Budhi,

Yes, from your description, such frequent database querty in each page
request is a potential cause though we can not make sure from the overview.
Also, the problem you met is not always reproduceable under test
environement (just occurs specifically in production environment...). I'm
afraid this will be really hard to narrow down and troubleshooting. At
least , we need to narrow down the problem to some certain page, so that
it'll make it a bit eaiser to reproduce...

Regards,

Steven Cheng
Microsoft Online Support

Get Secure! www.microsoft.com/security
(This posting is provided "AS IS", with no warranties, and confers no
rights.)

Nov 24 '05 #18

This discussion thread is closed

Replies have been disabled for this discussion.

Similar topics

10 posts views Thread by **ham | last post: by
115 posts views Thread by Mark Shelor | last post: by
13 posts views Thread by bjarne | last post: by
7 posts views Thread by James | last post: by
5 posts views Thread by Varangian | last post: by
By using this site, you agree to our Privacy Policy and Terms of Use.