I have an ASP.net Core application thats setup to use Environment Variables Configuration, I have published the app and created a Web Application thats running "below" the IIS Default Web Site.
The Web Application is running on its own App Pool and LoadUserProfile is set to true, and the identity is set to ApplicationPoolIdentity.
My problem is that as soon as I change one of my environment variables and recycle the app pool, the web application does not pick up the changed values of my environment variables (which it should).
To get the environment variables to get picked up by the application I need to run iisreset (just restarting the IIS from the GUI manager isnt enough)
However, I have the very same setup in a different server, with one exception, the Identity of the App Pool is set to a specific account, and in this scenario the application does indeed pickup any changes made to an environment variable after an app pool recycle.
Has anyone had the same issue? How did you resolve it?
It isnt ideal to be "forced" to run iisreset if your running multiple Web Sites and/or Web Applications in the same IIS.. so thats not really an option. But as mentioned it seems to have something to do with which identity is set for the app pool.. but I might be wrong.. anyone has any thoughts?
I have googled for a solution, however.. the only thing I did find was an issue on github thats closed by now where they solved it by running iisreset.. also alot of posts on the web mentioned to set LoadUserProfile to true for the application pool and also to recycle the pool, but as mentioned.. thats the thing that doesnt work.
please check the environment variables on IIS.
Open the IIS Manager and select your website
Open "Configuration Editor" from "Management" section
On the first dropdown (section), select
"system.webServer/aspNetCore"
On the second (from), select web.config
Then select "environmentVariables",you can add/remove buttons to configure as you want.
Related
We've been finalizing NopCommerce .Net Core web app which has been running great on a test server. I'm now trying to transfer the app to our production server, which did not have .Net Core. I installed the latest .Net Core hosting bundle and rebooted the server. I also have Web Deploy running on both the host and the client. I exported the app from the test machine and imported it into a newly created IIS site. After setting up the bindings - and enabling stdoutlogging, I try to see what's working, and get indication that "An error occurred while starting the application". No indication what the error is. Logs is not being written to. The event viewer tells me that:
Application 'MACHINE/WEBROOT/APPHOST/NOPCOMMERCE' started process '6980' successfully and is listening on port '41573', which is a random port not binded to.
One interesting thing I noticed on the test server is a "user" called nopCommerce which has full rights to the nopCommerce folder in inetpub/wwwroot. However this user does not show when I look at local users and groups. I am not sure therefore what this "user" is and if/how I should create it. Based on some advice from somewhere I temporarily gave everyone full rights to the nopCommerce folder, but that didn't work either.
Can anybody please aim me in the right direction?
Problem was a bad setup - access rights to subfolders of nopCommerce was one, which I solved by giving the users group modify rights. This might be a bad idea and I will do some more research. The other fault was a bad database login in the connectionString.
Ultimately I had to learn that instead of starting the app via IIS, it can be run from the command-line, and then messages and errors will be displayed in the DOS box. What to run is determined from the
I still don't know where the nopCommerce user comes from on the staging server.
I have a Asp.Net Core MVC application running on Windows Server 2008 R2 with IIS. But every time I update this application, I need to manually stop the applicationPool in IIS, and restart the applicationPool after I finish updating the app. Otherwise it will tel me "the xxx.dll is in use by other progress".
Is there any way to make this process easier?
A workaround For Windows with no down time and I am regularly using is:
Rename running .NET core application dll to filename.dll.backup
Upload the new .dll (web application is available and serving the requests while file is being uploaded)
Once upload is complete recycle the Application Pool. Either Requires RDP Access to server or function to recycle application pool in your hosting control panel.
IIS overlaps the app pool when recycling so there usually isn’t any downtime during a recycle. So requests still come in without every knowing the app pool has been recycled and the requests are served seamlessly with no downtime.
I am still searching for more better method than this..!! :)
January 2022 - for Linux
For Linux, we use Openresty nginx with Lua script to hold (sleep) incoming requests for few seconds until the service running .NET Core or .NET 5 or .Net 6 application restarts and then we release the threads we had hold.
Refer: https://github.com/basecamp/intermission
Finally I found my anwser:
I just need add a file named app_offline.htm to the IIS web root(not your project wwwroot folder), and remove it after you replace all of your file.
due to this issue you may need try both App_Offline.htm or app_offline.htm .
and this will allow you to use FTP client to update
Opening the web.config file in an editor and saving it will cause the web application to reload, even if you don't change anything. All DLLs should be replaceable, until a user hits the site, causing the web application to start again. I sometimes use that as a workaround.
A more full fledged solution would be to use Web Deploy, either through Visual Studio or by command line. This can take a litte while to set up, but offers more automation.
https://learn.microsoft.com/en-us/aspnet/core/publishing/iis#deploy-the-application-1
There is no way to hotswap in place DLL's.
Your best bet is to deploy to a new folder each time (For example a versioned folder), and change the website directory in IIS once you have fully copied your website onto the server.
I want to deploy on IIS my web site but I do not want to take whole project. I just need to take .dll file. Is their any way to do so.
I do not want to use visual studio only .dll file from the project to deploy.
The basic steps for deploying to IIS on windows server are as follows:
log onto the machine that is or will be hosting your application.
Use IIS Manager to create a new website for your application.
Create a new application in that site. I believe this also will automatically create an application pool with the same name for you and use it by default.
Specify the virtual directory for your application. This is going to tell IIS where to look for your mvc application. For this case lets assume it is C:\myApp
On your own machine Build the application however you build it with the correct solution configuration (i.e. Release mode). Let say the result of your build is located at C:\MyProject\bin
Copy C:\MyProject\bin from your machine onto your hosting machine at C:\myApp
You should be able to search these steps and find a step by step guide of how to accomplish them. Here is a link to some info on what sites, applications and app pools are to help you better understand.
http://www.iis.net/learn/get-started/planning-your-iis-architecture/understanding-sites-applications-and-virtual-directories-on-iis
Based on your sites requirements there will be some additional steps to set up security and alter bindings if you need to change them.
You don't need to deploy your entire website if you only make a change in a single assembly. You could copy the .DLL assembly directly to the bin folder of your website. This will trigger the Application Pool to be recycled in IIS and the changes will be taken into effect on the next request.
Very frustrated with all of this, hoping someone can assist.
I had a Silverlight application and WCF working together without issue for a year. In order to get them working, I had some pain initially but finally worked through it with help. All of the pain came from configuration/security, 401's, cross-domain hell, etc.
The way I have everything setup is that I have a WCF service that resides in it's own application/directory and runs in its own application pool.
On the same web server (IIS7), I have another application/directory with the Silverlight application that points to the aforementioned service.
The server name (for this exercise) is WEBSERVER1. We've created a CNAME for it that is WEB1. In the past, if the user went to http://WEB1/MyApp/ or http://WEBSERVER1/MyApp/ it would work. Suddenly yesterday it started behaving badly. Normal users started getting the Windows challenge/response prompt (and even if they entered the info they would get a 401 error).
My WCF service runs in a site that enables anonymous access (and this has always worked).
My Silverlight application runs in a site that has windows integrated (and this has always worked), since I capture the Windows username when they connect.
For the record, I did create a NEW application pool yesterday with an ASP.NET application that runs in it. This seems to work fine, but there is a chance creating this new application pool and application/directory has caused something to change.
I have a clientaccesspolicy.xml in my wwwroot folder, as well as in the folder for each of the two applications above (just in case). I have tried to promote NTLM over Negotiate as a provider (as that worked for another issue I was having on another server).
After trying some changes, I can't even get the thing to behave the same each time I call it. Sometimes it will prompt me for credentials. Other times it will work, but then say it failed to connect with the WCF service with a "not found". Other times it will actually work fine, but only if I am using the actual server name and not the CNAME. When using the CNAME I always get the crossdomain error, even though I have the cross-domain xml files in every directory root.
This is a nightmare, and makes advanced algorithm analysis seem fun and easy by comparison. Did Microsoft realize how difficult they made this combination of (IIS7/WCF/Silverlight/providers/permissions/cryptic or missing error messages) to get to work??
I found a solution that appears to be working.
In this case, I had to change the authentication mode for the default web site (which hosted the clientaccesspolicy.xml file) from anonymous access to Windows Integrated. I don't understand why this worked for a year or so and then stopped, but it seems to have resolved it.
The new application that I had deployed yesterday was a standard ASP.NET web application, which I put in it's own application directory and it's own application pool, to ensure that it would not cause this sort of issue. I'm still not even sure if it did.
The way I resolved it was by trying to navigate from my PC to the actual http://servername/clientaccesspolicy.xml file, and that was giving me a 401 error. I switched from anonymous to windows integrated on that default website (which has nothing in it except for that xml file) and that resolved the permission issue. I then had to permission the actual AD groups to have read access to that folder (if not they got the user/pw prompt and could not get through).
I know I've had this problem when I started working with Silverlight, but I can't for the life of me remember how to fix it.
I created a new RIA service application using the standard tutorial, added a table from the database and added a grid to display the results. Works great. Now I pull open the Web properties and change the web project to "use local IIS Web server". Suddenly the application will load up and give me the friendly "NotFound" error.
Please, someone remind me what I'm missing here.
I ran into this problem recently, and resolved it with help from this post on the silverlight.net forums.
Basically, I had Windows authentication and annonymous access enabled at the same time, and I need to disable windows authentication and restart IIS.
John
Not sure what the missing part is but I always start with Fiddler as will show the messages going across the wire. The actual messages can contains far more useful that the browser is hiding from the Silverlight plugin