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.
Related
What happens?
I publish a website to UnoEuro on a windows server.
When I do I keep getting the error message:
"HTTP Error 500.19 - Internal Server Error
The requested page cannot be accessed because the related configuration data for the page is invalid."
What have I set up/tried?
What I've done is that I've created a .NET Core 2.1 Web API which hosts fine locally on my computer using IIS. I have then published the web api to the hosting site UnoEuro where I can see that the files are uploaded as they should be.
I asked the hosting sites support, they told me to just ensure that it was deployed as self-contained but that otherwise they could not help further. It looks like there is a hosting issue of some sort related to that it cannot seem to find the webconfig file. At least that's what the error message says.
Is there any form of setup I should make? A configuration I should change? Should I set up some SSL settings in the web API which I might have missed? When I started the project I told it to enforce HTTPS, but surely I should get another error message if that was the fault?
I've tried publishing with the target runtime set to both x64 and x86.
I've never published an API like this to anything but Azure before, so I might be doing a thing or two wrong but I cannot see what it should be? I have Googled around but not found anything that seems relevant which has worked.
Most of the solutions seem related to changing the server environment which I don't have so much control over, I mostly just have control over the application/web api, at least that's what it looks like to me.
Hope you have any suggestions, I'm all eyes n ears.
Side question: Should I use IIS Manager for this? I'm currently publishing through Visual Studio 2017.
Images for further understanding
Publishing settings
Deployed/published files
Error message I get
EDIT:
I tried moving the files to /public_html on the same domain as ISS Manager seemed to indicate it expected the config file to be there.
That then resulted in a different error message as you see below.
Link from error message: https://go.microsoft.com/fwlink/?LinkID=808681
It's hard to say exactly, but I think the issue is that the ASP.NET Core hosting bundle is not installed on the server. There's two pieces to hosting in IIS. First, you need the .NET Core runtime, and then you need the ASP.NET Core hosting module for IIS. Deploying as self-contained buys you the runtime, but not the hosting module in IIS. There's non-standard sections in the Web.config created by publishing an ASP.NET Core app. The hosting module shims in support in IIS for these config sections (among other things). Therefore, the config error seems to imply that this is not happening.
If you don't control the server, there's not much you can do about it. Honestly, at this point, I wouldn't host anywhere where you do not at least have some measure of control over the environment. You can get a full VPS from numerous hosting services for $5 or less a month. Either continue hosting in Azure, or look into other respected cloud providers like DigitalOcean, AWS, etc.
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.
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'm stumped. There are several threads out there that reference a similar problem, but not the same one. I'm trying to get a simple WCF service up and running in Azure, but my problem is happening before I even deploy. I'm following this article:
https://azure.microsoft.com/en-us/documentation/articles/cloud-services-how-to-create-deploy/
Which says to create your project, build the "Hello World" service interface and class, and then start your project locally. I get a 403 error when I try to start my project.
The only step I'm not following exactly is that I don't have a "Start Without Debugging" option (I'm using Visual Studio Express Web Developer), so I just push "Start Debugging."
Why would I get a 403 error on my own machine? I am running Visual Studio as Administrator.
On the WebRole project right click on your .svc (Service1.svc) and choose "Set As Start Page" And then run the project. Make sure to do this on the WebRole project not the Azure Cloude Project.
Is the Azure project the default project or is the WCF Web Role? If it's the role itself, that may be your issue - it's trying to access the root of virtual directory containing the web service versus running the cloud app in the local emulator. If you browse to http://127.0.0.1:81/Service1.svc in a new browser window do you still get the 403?
It seems VS has lost the its project settings. Bring up "Properties" page on the web service in VS. Go to the "Debug" tab. Under Servers/Virtual Path: my configuration was set to just "/". Put the name of your server back in there "/Service1.svc" and it should work from the debugger now. This fixed the problem for me.
The problem went away when I upgraded to VS 2010 Pro. I never fixed it on Express, but after upgrading the problem was gone ... I had the "Start without Debugging" option in the Pro edition, so I suspect that the problem may have been specific to Express.
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