Good day everyone,
While working with Wix I am having a hard time solving the following error message:
Prompt for source of container: WixAttachedContainer
The scenario:
We have a setup project with the default Wix template that acts as a bundle for several packages, called MyBundle.exe.
We have a managed bootstrapper application wix exe with WPF and C# that acts as a standalone installer, but this needs to be included in the bundle stated above, called MyApp.exe.
MyApp.exe gets installed correctly when installing MyBundle.exe.
When uninstalling MyBundle.exe and having a DetectCondition to uninstall MyApp.exe. MyApp.exe gets removed correctly.
The Problem:
There is a scenario where MyBundle.exe will get removed but has to leave MyApp.exe installed behind.
After removing MyBundle.exe, and trying to remove MyApp.exe I get the following error in the log:
Prompt for source of container: WixAttachedContainer, path: "Path_To_EXE"
If I install MyApp.exe as a standalone and uninstall it, everything works just fine.
If I install MyBundle.exe (MyApp.exe included in MyBundle.exe) and uninstall MyApp.exe before removing MyBundle.exe, this works fine as well.
I have read the logs of MyApp.exe where the error comes up and I found that MyApp.exe is trying to reach the cache folder of MyBundle.exe
Prompt for source of container: WixAttachedContainer, path: C:\ProgramData\Package Cache\{Bundle_GUID}\Dependencies\MyApp.exe
But at this point, that folder is no more on the system.
The configuration of the ExePackage in MyBundle.exe is as follows:
<ExePackage Id="MyApp_Package"
Description="MyApp Package"
DisplayName="MyApp Package"
SourceFile="My_Path_To_Package"
Compressed="yes"
Permanent="yes"
PerMachine="yes"
Vital="yes"
Cache="yes">
</ExePackage>
I have tried several combinations of the tags Permanent, Cache, but still to no effect.
I have tried already hooking up to the ResolveSource method inside the C# code on MyApp.exe following instructions like this:
Managed Bootstrapper Application "Failed to resolve source for file" but to no success.
Can any of you point me in the correct direction?
UPDATE: Both MyBundle.exe and MyApp.exe, as well as the msi included are signed using signtool.
Thanks.
Take a look at "insignia"
https://wixtoolset.org/documentation/manual/v3/overview/insignia.html
It's not enough to just use signtool to sign the boostrapper EXE. The bundle inside it needs to be signed also.
After upgrading my project to ASP.NET Core 2.2, I tried to run the application (locally of course) and the browser displayed an error message like in the below screenshot.
no more errors notified by visual studio error explorer. I don't know what's happen.
In my case, I upgraded some nuget packages to net core 2.2, but I did not have the net core 2.2 sdk installed, so I went to net core website to download the latest sdk or runtime package, and then I did a net stop was /y and then a net start w3svc in the CMD as administrator. Problem solved for me.
I encountered this error after trying to publish from VS2017 to the production Windows 2016 server. (It worked fine in IIS Express on my local Win10 PC.)
I updated packages, all versions matching and updated in my code, .net core versions matching, restart IIS, rebooting... no joy.
In the Publish > Configure > Settings (left tab) I had to set the Target-runtime from "Portable" to "win-x64" (or whatever is relevant to your environment). I also opted to "Remove additional files at destination."
"Portable" is the default setting. I'm not sure what it takes for the "Portable" runtime to work properly, but might save someone else some time if a "Portable" runtime is not something you need.
Generally speaking, I get this error if something is mismatched in my environment. For example, one time I was upgrading one of my projects to .Net Core 3.1 from 2.2 and hadn't installed the ASP.NET Core Runtime Hosting Bundle on my server:
https://dotnet.microsoft.com/download/dotnet-core/3.1
Also, you can get this error if your Application Pool is set to True for Enable 32-Bit Applications. Try:
IIS Manager > Application Pools > app pool name > (right click) Advanced
Settings > Enable 32-Bit Applications = False
I ran into this issue and had a different solution. For me it was that I had a package that was out of date with the application (I had updated it on NuGet, and the library hadn't been replaced in production). Updating the package fixed it for me.
Note with this: I had to manually run dotnet.exe with the project dll in order to see the message that fixed it for me.
Hope this helps someone else down the road.
you have 2 solution(this answer works on windows server I do not know anything about linux server).
first:
copy all folder(except bin and obj folder) of your project to server
open cmd in your project folder then run this command: dotnet run then all warning and error show to you(if you have error about above command not recognize download dot net core sdk from this link)
second:
you must changed hostingModel attribute from OutOfProcess to
inprocess in web.config and you can change stdoutLogEnabled to true
value for get your project error in logs folder
read your projects errors and fix those.
in my case web.config is:
<?xml version="1.0" encoding="utf-8"?>
<configuration>
<location path="." inheritInChildApplications="false">
<system.webServer>
<handlers>
<add name="aspNetCore" path="*" verb="*" modules="AspNetCoreModule" resourceType="Unspecified" />
</handlers>
<aspNetCore processPath="dotnet" arguments=".\BMS.dll" stdoutLogEnabled="false" stdoutLogFile=".\logs\stdout" hostingModel="OutOfProcess" />
</system.webServer>
</location>
</configuration>
and I change it to:
<?xml version="1.0" encoding="utf-8"?>
<configuration>
<location path="." inheritInChildApplications="false">
<system.webServer>
<handlers>
<add name="aspNetCore" path="*" verb="*" modules="AspNetCoreModuleV2" resourceType="Unspecified" />
</handlers>
<aspNetCore processPath="dotnet" arguments=".\BMS.dll" stdoutLogEnabled="true" stdoutLogFile=".\logs\stdout" hostingModel="inprocess" />
</system.webServer>
</location>
</configuration>
In my case it was the log level set incorrectly in the appsettings.json. Instead of Warning I've had Warn and this crashed the app with above error.
Seems that everyone has a different answer for this. I also had this issue as well. There are many different things as you can tell that cause this issue. If you don't find any of these solutions helpful or have issues trying to go through all these different solutions, you can try running your application from the command line from the publish folder.
After publish, if you receive this error, go to your publish folder, and then open a command/terminal window, after that type dotnet .\YourStartupProject.dll, you should receive an exception error, which should make fixing the issue easier.
For example, this is an error I received on trying on a new environment without setting up a SQL server, and of course, would receive this error.
Application startup exception: System.Exception: Could not resolve a service of type
'YourStartupProject.DataServices.DbContext.DbContext' for the parameter
'context' of method 'Configure' on type 'YourStartupProject.Startup'. --->
System.ArgumentNullException: Value cannot be null.
Parameter name: connectionString
Once you resolve your error, try it again, rinse, repeat.
For me the issue was caused by dotnet publish creating a web.config entry stdoutLogFile=".\logs\stdout". The correct value should be stdoutLogFile="\\?\%home%\LogFiles\stdout".
MSDN reference: https://blogs.msdn.microsoft.com/waws/2018/06/10/troubleshooting-http-502-5-startup-issues-in-azure-appservice-for-asp-net-core-websites/
This could be a bug in the ASP.NET Core 2.2.0 runtime which may have been fixed in a later version.
Follow this steps:
create a directory in root of your project : logs/stdout
open the web.config file from root of your project and find this line:
<aspNetCore processPath=".\web.exe" stdoutLogEnabled="false" stdoutLogFile=".\logs\stdout" />
set stdoutLogEnabled as true and save it
reload your app and see the logs in the directory : logs/stdout
I ran into this issue today with my hosting - locally everything is ok but once I publish, I get this error.
I looked through the packages and found out that some .net core stuff was upgraded to 3.0 preview.
Then I changed the build option in VS2019 from "Framework-Dependent" to "Self-contained". It took 5 times longer to build and publish but now it works.
Now I'm checking with host tech support what might be an issue - officially they support 2.1 / 2.2 only, so this might be these packages from 3.0 Preview, however target build is 2.2.
My issues was malformed appsetttings.json file. I enabled standard out logging via web.config and was able to get the underlying exception throwing this error.
If resetting the project and manually copying Program and Startup classes worked for you, then something was clearly messed up. There are some bigger underlying problems with this. Using the OutOfProcess hosting model is okay, but with .Net Core 2.2 you should be able to use the InProcess hosting model, since it is naturally faster: everything is processed in IIS, without an extra HTTP-hop between IIS and your app's Kestrel server.
If you right-click your project file in the visual studio solution explorer, please make sure that AspNetCoreModuleName tag has AspNetCoreModuleV2 value (as opposed to the older AspNetCoreModule). Also, examine Windows Application Event Log to identify the potential culprit. Even though error messages there are somewhat cryptic, they might point you to the exact line number in the code that caused the failure.
Finally, in case you use CI/CD with TFS, there may be environment variable(s) in appsettings.json file that were not properly replaced with the actual values (URLs, etc.).
Looks like i had the same issue. It's happens because if you don't have global.json file in solution, then VS build(publish) .net core app with the last version that installed on your pc. So, i do the next solution:
add a global.json file with .net core version.
{
"sdk": {
"version": "2.2.402"
}
}
From learn.microsoft.com:
global.json can be placed anywhere in the file hierarchy. The CLI searches upward from the project directory for the first global.json it finds. You control which projects a given global.json applies to by its place in the file system. The .NET CLI searches for a global.json file iteratively navigating the path upward from the current working directory. The first global.json file found specifies the version used. If that version is installed, that version is used. If the SDK specified in the global.json is not found, the .NET CLI rolls forward to the latest SDK installed. Roll-forward is the same as the default behavior, when no global.json file is found.
https://learn.microsoft.com/en-us/dotnet/core/versions/selection
This happened to me when I deployed code using Entity Framework Core with migrations, and there was mismatch between the state of the database and the migrations in the code.
This happened to me first time publishing an Azure Web App. Here is how I solved it:
Browse the site using Kudo/FTP. In the root folder there is a LogFiles folder where you find eventlog.xml. In this file I could see that my web app had an SqlException when Entity Framework Core was trying to setup the database, which lead me to check the database permissions (which was the problem for me).
This is what worked for me:
- I ran the startup file of the project in the deployed (IIS) folder. Note that: this will not solve the problem but will inform you about what the problem is. In my case, the cause of the problem was a database migration that failed
Another answer that might help other people in the same case: we have an AppService on Azure where there are 3 NETCore project deployed on 3 different path:
One for Web (/webapi)
One for Mobile (/mobileapi)
One for Functions serverless, in our case was it was AzureFunctions (/functionapi)
Since the upgrade to NETCore3.x, we understood that the hosting model by default was "In-Process" so we had to edit the .csproj file to explictly set the hosting model to "Out-Of-Process" like this:
<PropertyGroup>
<TargetFramework>netcoreapp3.1</TargetFramework>
<AspNetCoreHostingModel>OutOfProcess</AspNetCoreHostingModel>
</PropertyGroup>
But it was not enough: in fact, we also have to edit Program.cs. Why ? Because in Program.cs the one generated by default in NETCore3.x you have the following code:
public static IHostBuilder CreateHostBuilder(string[] args)
{
return Host.CreateDefaultBuilder(args)
.ConfigureWebHostDefaults(webBuilder =>
{
webBuilder.UseStartup<Startup>();
webBuilder.ConfigureKestrel(o => o.AddServerHeader = false);
});
}
When we replaced this by the old code by NETCore2.x version like below:
public static IWebHostBuilder CreateWebHostBuilder(string[] args)
{
return WebHost.CreateDefaultBuilder(args)
.UseKestrel(options => options.AddServerHeader = false)
.UseStartup<Startup>();
}
After deployment, the error 502.5 ANCM Startup Failure was gone :) Hope this answer can help other people.
BTW I know this post is related to NETCore2.2, we also met the same problem but we decided to switch to NETCore3.1 because NETCore2.2 was no more supported and this version was also buggy on some other points.
My .NET Core site was worked fine, but after a while, I got this error (HTTP Error 502.5 - ANCM Out-Of-Process Startup Failure ...);
I tried different methods. Finally I Add new web site in IIS (with other port), then the error was solved.
I got this issue in ASP.NET Core 2.2 project and its resolved for me just by Clean and Rebuild project.
This error started appearing on our Dev server. I had been using this publish command which creates a "self-contained" folder of files for deployment.
dotnet publish -c release -r win7-x64 --output:bin/self_contained
My fix was to instead publish a "framework-dependent" deployment using the following command:
dotnet publish --output:bin/framework_dependent
The dev server did have a few versions of .NET Core installed (2.2.3 and 2.2.5) in this folder *C:\Program Files\dotnet\shared
I am still not clear on why the self contained publish does not work. You might think the self contained publish would be the more reliable method, but in my case it was not.
This .NET Core blog post was helpful.
I got this same error while deploying .Net core app which was targeting .Net framework on Windows server. I checked event viewer on the server and turns out server didn't have .net 4.7.2 installed.
Installing it resolved issue for me.
Yet another scenario that caused this issue for me:
I am running the app pool identity with a service account and I had to run dotnet dev-certs https under this user to get rid of "System.InvalidOperationException: Unable to configure HTTPS endpoint." during startup.
Be carefull publishing.
When i publish it to my PreProd envitoment this conf works well:
Portable
But on my Prod enviroment that conf does not work. I had to choose the especificated one:
win-x64
I dont know the reason about that. If someone know i'll gratefull to know!
The problem occurs when I try to deploy the asp.net core (out-of-process hosting model) website to windows server 2012r2 IIS in production env.
I fixed this with this solution:
Change application pool identity to administrator.
Same failture happent on project publish.
The issue ralated with the latest Microsoft.AspNetCore.App package. Just downcast it from from 2.2.x to 2.2.0
or goto dotnet.microsoft.com/download/dotnet-core/2.2 and get newest dotnet-hosting installer
I was also getting the same issue. And when I looked at the Output window of my solution.
Then I was able to see a different error, which is "The target process exited without raising CoreCLR started event", to fix this I had to remove the Microsoft.AspNetCore.All from my Nuget Packages and install Microsoft.AspNetCore.App. I also had to install the correct .Net SDK from here. Once this is done, restarted my machine and open the solution, the error was gone. Hope it helps
If you are working with ASP.Net Core version 2.2 then in appsettings.json just comment the line -
"AllowedHosts": "*"
it resolves the issue. My application working fine.
This error can be happened because of many reasons. In my case it was an exception due to invalid format of appsettings.json . How I found out is by enabling stdout log in web.config.
For me the issue was a missing appsettings.json
I select the appropriate appsettings.json file (appsettings.production.json or appsettings.development.json) based on an environment variable. Turns out the appsettings.json is required even if you dont use it.
My problem was with the web.config file after publishing. The processPath in the aspNetCore tag was missing the file extension. In my case it was .exe
In my case EF Migrations thrown exception about blocking executing one of them because of a potential data loss.
I had to look into custom app logs (most often Log folder) to find out that.
I guess the Error mentioned in the Question is due to problems during app start stage. And indeed the migrations are run during starting an app, so if they fail the app is not able to complete starting.
So in general when we get such Error we should focus on things that impact on starting logic of the app.
A collegue is trying to install a nuget package into a simple default c# web application. It fails almost instantly.
Is there an argument I can provide to Install-Package <some nuget package> in the Visual Studio Package Manager Console to get some verbose information to help debug why the installation fails?
Error Message:
An error occurred while retrieving package metadata for '' from source 'MyGet'.
Info:
Visual Studio: V2015
NuGet extension: 3.4.4.1321
Nuget package source: MyGet
Sample NuGet.config file found in the root directory of the solution:
<?xml version="1.0" encoding="utf-8"?>
<configuration>
<packageSources>
<add key="nuget.org" value="https://api.nuget.org/v3/index.json" protocolVersion="3" />
<add key="MyGet" value="https://www.myget.org/F/<our package>/api/v2" />
</packageSources>
</configuration>
For myself, I can install the package fine. In fact, we have 5 packages in this MyGet public repo and I just installed 2 of the packages, just then .. when I test this out (again) before I created this SO question.
Anyone have a suggestion, please?
UPDATE
As stated above, this is using the PACKAGE MANAGER CONSOLE, not the CLI.
Using the -verbosity detailed in the PMC this is what happens..
PM> install-package xunit -verbosity detailed
Install-Package : A parameter cannot be found that matches parameter name 'verbosity'.
At line:1 char:23
+ install-package xunit -verbosity detailed
+ ~~~~~~~~~~
+ CategoryInfo : InvalidArgument: (:) [Install-Package], ParameterBindingException
+ FullyQualifiedErrorId : NamedParameterNotFound,NuGet.PackageManagement.PowerShellCmdlets.InstallPackageCommand
You could try adding the -Verbose parameter to your PowerShell command.
install-package xunit -verbose
You can also try looking at the $error object to see if that has more information, such as an exception callstack.
$error.Exception.StackTrace
The above may or may not give you more information.
Looks like you are running into http://blog.myget.org/post/2016/02/21/Two-of-my-packages-are-treated-as-one-Help!.aspx. There is a 0.7.0-dev and a 0.7-dev version of the package on the feed, which NuGet treats as the same version.
The solution is to remove one of these two packages.
You can use the -Verbose switch in the PowerShell-enabled Package Manager Console to get more details.
To rule out client-connectivity issues to MyGet, can you try diagnosing client connectivity to the feed?
https://www.myget.org/F/<feedIdentifier>/api/v2
Try using Fiddler to retrieve a specific package from the feed on that user's machine using URL format:
https://www.myget.org/F/<feedIdentifier>/api/v2/package/<packageId>/<packageVersion>
If it is a private feed, you'll need to authenticate your request, e.g. using the pre-authenticated v2 endpoint:
https://www.myget.org/F/<feedId>/auth/<apiKey>/api/v2/package/<packageId>/<packageVersion>
Clearing the NuGet client's HTTP cache may also prove useful if something is corrupt in the cache. You can clear it using the NuGet commandline command:
nuget.exe locals http-cache -clear
If the package is already in local cache, then the NuGet client will resolve it from local cache instead. It may be that the package in local cache is corrupt, in which case you can use the following NuGet commandline command:
nuget.exe locals all -clear
Finally, it may also be worth looking at the nuget.config hierarchy. The most-local nuget.config will inherit config settings from higher up the chain (e.g. machine-wide settings), one of which is the <disabledPackageSources> element. To ensure this one is not acting up here, add the following to your most-local nuget.config and retry:
<disabledPackageSources>
<clear />
</disabledPackageSources>
If none of the above helps you in resolving the issue, feel free to use the nuget commandline (nuget.exe) with -verbosity detailed as it may provide more details, such as the actual HTTP requests being made.
Also, make sure you use the latest versions of the NuGet client tools (available here)
We have a private nuget repository for sharing internal components. For now, we are just using a network share.
I am running into some problems though:
Everything works fine if I use the Package Manager Dialog box -- I can search for my packages, install them, and build.
However, if I use the Package Manager Console command: install-package My.Package (where My.Package is the exact string in the <id> tag of the nuspec file, I get an error indicating the nuget could not find the package.
If I include the version number in install-package: install-package My.Package.1.0.0, then it works from the console.
BUT, our build server fails to download our packages with the same error I get from install-package. (our build server uses package restore)
I have compared my nuspec files with some publicly available packages that work find. The only difference I can find is that our nuget packages are on a network share, but public ones are on a web server.
Is there a behavior difference if you use a network share?
I figured out what the problem was -- it was mostly user error.
I went through my nuget config, and I found this:
<activePackageSource>
<add key="nuget.org" value="https://www.nuget.org/api/v2/" />
</activePackageSource>
That setting tells nuget to only search nuget.org, and not our private feed. Had to change it to this:
<activePackageSource>
<add key="All" value="(Aggregate source)" />
</activePackageSource>
I'm still not sure why it worked before (sometimes) if you included the version number (maybe it had to do with whether the package was already locally cached?? It seems as though it should not have worked at all until I made this change. This fixed it though.