Oracle Data Provider Pegs IIS Worker Process When Web Site Is Stopped?
May 6, 2010
We're experiencing a nasty issue in Oracle 11g Release 2 where the w3wp process takes over and entire processor core, and debugging shows that the Oracle data provider is throwing ThreadAbortExceptions infinitely. A developer found this issue by doing the following:
1) Browse a web site that uses Oracle data connections locally (http://localhost/OracleWebSite - we use IIS, not the ASP.NET dev server, for all of our sites). This ensures that the w3wp process is running and that an active Oracle pool of connections exists in the app pool.2) Stop the web site (or perform a Rebuild All operation in Visual Studio on the web site in question).Our Oracle connection handling in the affected applications (all Oracle web apps) is well-established and robust. This issue does not occur if we disable connection pooling. This issue does not occur in Oracle 11g Release 1.
I have a .NET infra code running both within the IIS worker process and within a desktop client app. How can the .NET code determine whether it is running within an IIS worker process?
I know that I could check the name of the process (w3wp.exe, for instance), but I would like a more robust approach. I wish to make a side note. This is not a production need. I need this information to enable certain scenarios useful during the development and testing phase. Specifically to ease the testing of secure vs non secure configurations.
Does ASP.NET worker process still return data in chunks of 31KbThis MSDN article written in May 2004 specifies the following.When using the ASP.NET process model,the ASP.NET worker process sendsresponses back to the client, it firstsends them through IIS in 31-kilobyte(KB) chunks. This applies to .NETFramework 1.1, but it could change infuture versions.
We have a site that is runnig II7 64 bit with the oracle .net 64 bit provider, when I try and open this in web developer 2010 express (I'm running windows 7 pro 64bit) it says it can't load the oracle dll as its an incorrect format. I got the impress that web developer 2010 was 64bit capable.
Ofcourse this is extremly annoying as I don't get any syntax formatting for the project or intellisense.
I'm using VS 2010. And in VS 2010 it's even more painful now to attach to the worker process when you want to test your local IIS site via IIS, not the VS built-in web server.It's more painful because now when you select the worker process () via the Debug Menu | Attach to Process | w3wp.exe it brings up an additional prompt asking if you're sure you want to do this.
I have increased the number of worker process of default application pool to 15(IIS->Application Pools->DefaultAppPool->Properties->Performance->webgarden).after increasing the worker process i am facing session timeouts
My application is using a state server session managemnet and a predefined machineKeyCan you please guide me in how to increase the worker process without session loss
In IIS application pool I have set 2 in maximum worker process in process model, will application object, session object and cache object will be shared by the all worker process or will all worker process have different application object, session object and cache object.
I'm in a need to get a list of all the ASP.NET running worker process and its associated application pools and process ID.Is there a way to do this pragmatically (C#) or a powershell script?
My web application is a product which has to deploy to a variety of web servers. Is there a way to determine the account name that the .Net worker process account is using at runtime?
First of all, I'm not a pro on IIS configuration topic.With that being said, I got a website written in VB.NET 3.5 Web Forms. I have a small web service, ASMX, running as a part of this website. One page on the website calls the web service with AJAX (jQuery $.ajax function, post). The server is running Windows 2003 and IIS 6.0. The website and the web service are sharing the same Application Pool.
What is the difference between IIS Worker process, Application pool and application domain? Do we have more than one application domains in one application pool?Does one application domain means one virtual directory? If we have 2 virtual directories pointing to same physical folder. Will they share the same memory or different when there instances are opened in the browser? Will they be in 2 different application domains?
We recently had to enable Memory Recycling on a 2.0 app pool (IIS6). We checked both boxes, leaving in the default values of 500 MB for virtual memory and 192 MB for used memory. Here's where it gets interesting. Every 60 seconds, on the dot, the worker process is forced to recycle.Our tests were spamming the site with concurrent users trying to get the virtual memory up to 500MB, but we never could (as viewed in task manager). However, that didn't matter. We tested it by simply recycling the pool and hitting one page, once, and seeing around 30MB allocated and then just waiting 60 seconds and watching the process die in task manager (and a new one start up).We've increased the value for maximum virtual memory to 768MB and the issue *appears* to have gone away, but just arbitrarily accepting a fix isn't how we like to do things. I haven't found any officially reported bugs, kb articles, or solutions on blogs that point to what is causing this. To me, this means either it's some crazy environmental issue or the answer is really obvious and we just can't see it.
I made a small app to capture frame from video using asp.net, in visual studio, it works fine. however the problem arises when i try to use any other servers, I am currently using iis 7.5, and trying to debug on server, but alas when i reach some region of code the attached w3wp.exe just get detached, the process crashs, no debug available anymore.
In my application i am sending emails of bigger size (> 2.5 Mb). Now till the emails are being sent, the application is in hanged up state (mouse cursor is of busy state).One of my friend suggested to use background worker that "could" display message something like "Please wait...Sending email". I have no idea as to how to use the background worker process. if there is any other quicker/easier alternative,
The architecture of the application is straight forward. There is a web application which maintain account holder data. This data is processed and the status of account holders is updated based on number of business rules. This process is initiated using a button on the page and is a long running process (say 15 mins). A component is developed to do this data processing which internally calls stored procedures. Most of the business rules are kept in stored procedure.
we're hosting a webservice that runs with up to 9 worker processes in it's application pool. There's a memory limit configured for the application pool to keep the system/server performant. Now my concern is that, if the memory limit is reached, a worker process may be recycled although there is an active request handled by this worker process that needs to be responded/finished first. The problem is especially with purchase/order requests, as such requests typically run several minutes and must not be aborted by any worker process recycle process. Is there any configuration setting that allows me app pool recycling by granting that no active process will be abandoned?
Does anybody know how I can trigger the ASPNET_Worker process on Windows 2000 and 2003 servers? Apparently, every time we reboot our servers on Saturday nights our monitoring software generates alerts for the ASPNET_Worker process and they don't go away until during the morning hours. This leaves me to believe that this process doesn't automatically run after a reboot and only starts to kick in when clients access our websites during the day. I'd like to be able to trigger the process after a server reboot so our monitoring software doesn't identify it as down.I was reading this post on http://weblogs.asp.net/jgalloway/archive/2006/06/01/Avoid-IISRESET-in-ASP.NET-applications-_2800_added-bonus_3A00_-ASPRESET_2900_.aspx and I believe the code on that article is what I need. However, I am only a SysAdmin and have very little knowledge about programming. The writer mentioned compiling the code using SnippetCompiler but after downloading the software I can't seem to figure out how to turn the code into an .exe.
I have a solution that was originally created using VS 2008. I have quite a few unit tests that all passed when ran in VS 2008.
Yesterday I opened this solution in VS 2010 and converted it using the wizard. The solution built just fine, it runs just fine, BUT now all tests that access the db fail. I get the same error on every test (below).
"The runtime has encountered a fatal error. The address of the error was at 0x5b3b5ad7, on thread 0x19a4. The error code is 0xc0000005. This error may be a bug in the CLR or in the unsafe or non-verifiable portions of user code. Common sources of this bug include user marshaling errors for COM-interop or PInvoke, which may corrupt the stack."
In test classes that use ClassInitialize() and ClassCleanup() methods, all the tests fail. All tests fail with the same error, "The agent process was stopped while the test was running." Tests in test classes that do not utilize ClassInitialize() and ClassCleanup() (these tests do not hit the database) all pass.
When this was a VS2008 project all 153 tests passed. No changes have been made to these tests, other than converting the project to a VS2010 project.
I've been playing around with one test class. I commented out the ClassInitialize() and ClassCleanup() methods. I rewrote one test method and mocked everything. I tried running just that one test. I still got the "The agent process was stopped while the test was running" error.
Visual Studio 2010, version 10.0.30319.1 RTMRel
Test framework: integrated VS 2010 MS Test
Mocking framework: Not using mocking, I have a db that is destroyed and recreated in MyClassInitialize(), it is populated with known values. Yes, I know I should be using mocking, but this is an older app and it has not been updated with all of the new bells and whistles yet.
DB Framework: NHibernate
Target Framework is 3.5 for all projects in this solution except for the test project. The target framework is 4.0 for the test project. I tried changing it to 3.5, it will not let me.
"Attempted re-targeting of the project has been canceled. You cannot change the specified .NET framework or profile for a test project."
I've run into a particularly nasty bug where an ASP.NET website attempts to make a call to a WCF service method that sends IIS into a death spiral...that ultimately brings down the associated app pool. It never gets to our log4net code that outputs unhandled exceptions.I was watching w3wp.exe processes spin up and die, so I figured I'd take a dump file on termination using ProcDump from SysInternals using the following command
Unable to start debugging on the web server. The IIS worker process for the launched URL is not currently running.I'm running Windows 7 64 bit, VS 2008 sp1.When I first try to debug I get the error msg above. I wait a few minutes (presumably some service is starting) and debugging will start but no breakpoints are recognized.Some of the things I've tried:Changing the binding of my site from a static IP to "*".
Enabled windows authentication.made sure localhost = 127.0.0.1 in the hosts file.The website was built under windows xp and whatever version of IIS ships with xp. My web config has no errors that I know of but might be missing something required for iis 7 (??)
This is a really odd situation, so hopefully I can explain it well enough.
I am deploying an ASP.NET 4 webforms application to a Windows Server 2003 SP2 server running IIS6.
Here's the problem -- when the application pool recycles its worker process (w3wp.exe), about 80% of the time, I will get an ReflectionTypeLoadException error trying to access any page in the app that contains an EntityDataSoure every time I try to view it.
However (this is the interesting part) -- the other 20%, it works just fine. I've actually resorted to turning off recycling the worker process entirely for this application pool and just add/remove whitespace from web.config forcing the site to recompile until I get a "good" w3wp.exe.
If this isn't clear, what I'm saying is: the actual worker process doesn't work at all for pages containing an EntityDataSource for about 4/5 times it starts, but still manages to serve all other pages just fine. Once you get a worker process that manages to serve a page with an EntityDataSource, it works every time until that process gets recycled.
My question is, how can I debug this? It works fine on my dev machine, it works fine on the server as long as you get a good process running, but iisreset or a server restart or anything that kills the worker process is almost guaranteed to cause the site to not come back up and throw this ReflectionTypeLoadException.
I have 2 sites on IIS, one is the live site and the other is a site that is only started when there is maintenance being carried out on the live site.In a deployment scenario I STOP the live site and the START the maintenance site so that users receive a friendly message advising of the upgrade.he only issue I have is when I start up the live site it obviously has to JIT and this can take up to 3 minutes.Is there anyway to have this JIT before I unleash it to the users?EDIT: Just to clarify, this site is a CMS, so the marked answer below works for me due to only having 1 page to compile.
I'd like to know if there are any good EF providers for Oracle out there, and more specifically what experiences people have had with them. We're looking for a provider that is either free or paid for, but I'm struggling a bit with finding any real opinions on what the different frameworks are like to work with in a production environment.
EDIT,I should clarify that we are using Oracle 11g.
