If I create a net core 2 console app and get it to reference another project e.g. MyLibrary.csproj
This (MyLibrary.csproj) is a net core class library
If I run dotnet publish -c release --output test1
then in the output folder their is a runtime folder present
I have not found anywhere that describes this folders purpose.
Any one have a link?
Also do I need to copy this as part of my deployment?
The example I have has a reference to System.Data.SqlClient.dll which is present in the root publish folder(test1) so why does it need to get it from the runtime folder when I try to run via dotnet my.dll?
From the docs (https://learn.microsoft.com/en-us/dotnet/core/tools/dotnet-publish?tabs=netcore2x)
dotnet publish - Packs the application and its dependencies into a folder for deployment to a hosting system.
Related
I want know what is the command line for making a directory a startup project.
I have 2 directories inside EmployeeManagement and on sln file:
Models
Web
Employeemanagement.sln
Now instead of Right clicking and making Web the startup project, is there a way of doing it using command line interface.
Thanks for your help.
In the .Net 5.0, you can do: dotnet run --project , everytime you need to run the project.
CLI Example:
dotnet run --project Web/Web.csproj
I have been trying to build the following project structure with dotnet cli on netcoreapp3.1:
project_root:
project.sln
projecct1
--------project1.csproj
project2
--------project2.csproj
third_party
--------third_party1
------------third_party1.csproj
The third party project is referenced in one of the projects and it is in its self a netsandard2.0 assembly
running dotnet build produces the following error:
error NETSDK1005: Assets file '...\third-party\third_party1\obj\project.assets.json' doesn't have a target for '.NETCoreApp,Version=v3.1'. Ensure that restore has run and that you have included 'netcoreapp3.1' in the TargetFrameworks for your project.
Building the third party on its own works fine.
I realize that the error stems from the structure of the project. Any ideas how I can build the project with getting this error?
Ok so answering my own question. The problem was not the structure. When I built the project in command line I use dotnet build -c Release -f netcoreapp3.1
Since the project is a mix of netcoreapp3.1 and standard2.0 assemblies the standard2.0 assemblies failed to compile. If your framework is set in the csproj file not need for -f apparently.
In visual studio solution I have single .net core 2.0 library project. And to the publish the library i am using dotnet publish -c release command
however its not zipping the publish folder. I have read the issue 6598 and use the suggested approach using dotnet build command as below
dotnet build ApiRouting.sln /nologo /p:PublishProfile=Release /p:PackageLocation="C:\temp\Routing\package" /p:OutDir="C:\temp\Routing\out" /p:DeployOnBuild=true /p:WebPublishMethod=Package /p:PackageAsSingleFile=true /maxcpucount:1 /p:platform="Any CPU" /p:configuration="Release" /p:DesktopBuildPackageLocation="C:\temp\Routing\package\package.zip"
`
but that did not work either.
My project is aws lambda project which is C# library project not asp.net web project so i tried removing /p:WebPublishMethod=Package option but that did not work either.
Questions
1>What parameters i need to pass to publish command so that it would create zip file of publish folder.
2>In linked issue 6598 why its suggested to use build command instead of publish when build command only builds the project?
(on side note i can use aws tools for visual studio and use Publish to AWS Lambda and it creates zip file and deploys it to AWS directly from visual studio. However, we are using Jenkins for CI so i want use dotnet cli to create zip file so jenkins can execute that command and create zip file.)
i found it. These 2 links helped me
https://docs.aws.amazon.com/lambda/latest/dg/lambda-dotnet-how-to-create-deployment-package.html
https://docs.aws.amazon.com/lambda/latest/dg/lambda-dotnet-coreclr-deployment-package.html
first installed Amazon.Lambda.Tools
dotnet tool install -g Amazon.Lambda.Tools
and then to package and deploy
dotnet lambda deploy-function apirouting –-function-role myrole --profile lambdadep --profile-location C:\test\testawsprofile
I'm attempting to create a script to simplify the process of publishing a .NET Core website. I'm running into an issue when I run dotnet publish against an already running server. The server is IIS with the dotnet bundle installed, so IIS uses its app pool to start dotnet.
Here's my batch file. I'm happy to use another script type:
cd src/app
dotnet build --no-incremental
dotnet publish --framework netcoreapp1.0 --configuration Release --output ../../dist
When I run the script I get this error:
"The process cannot access the file 'C:\inetpub\wwwroot\app\dist\app.dll' because it is being used by another process."
This makes sense, it appears I need to stop, deploy, and restart dotnet. Can I do this from the script? Or is my approach to this problem wrong?
The best way is to drop an app_offline.htm file to your application folder. This will make IIS stop your application and serve the contents of the app_offline.htm file to the user while you are copying the new version. Once you complete copying the new version of your application remove the app_offline.htm file and IIS will start your application.
You can find more details on running ASP.NET Core applications with IIS in my post.
Based on Pawel's answer, I have a deploy folder containing my app_offline.html file and multiple deploy scripts to IIS. Here's a sample script I use to deploy:
copy .\app_offline.htm C:\hosting\my-project\app_offline.htm
dotnet publish ../MyProject.csproj -r win-x64 -f netcoreapp2.1 --self-contained -c Release -o C:\hosting\my-project
del C:\hosting\my-project\app_offline.htm
I think this is a valid solution, but doesn't help when I want to script the build process.
Stop-Website "xxx"
Stop-WebAppPool "xxx"
Start-Sleep -Seconds 5
dotnet publish --output d:\publocation
Stop-WebAppPool "xxx"
Start-Website "xxx"
if you've created a published profile in Visual Studio and you're using IIS, then you can use that profile instead of writing directly to the destination directory:
dotnet publish /p:PublishProfile=Properties\PublishProfiles\IISProfile.pubxml
I have an ASP.NET core application that I've been auto-deploying to an azure app service on commit to a git repo. It worked fine as a project.json type project.
I've converted my project.json to myproject.fsproj and it builds and runs locally. On comitting the .fsproj to git, the deployment was triggered, but it failed with the activity log containing one line: 'D:\home\site\repository\myproject.fsproj' is not a deployable project.
I guess it's an issue with the default kudu deployment script? Does anyone know how to sort this out, or do I need to submit an issue/RFC to the kudu guys?
UPDATE
I generated the original .fsproj using:
dotnet new mvc --language f# --framework netcoreapp1.0
I've since made changes to it, so I will try to do a minimal case later tonight.
Turns out that the default deployment in azure wouldn't deal with this.
Following the answer on Kudu Deployment Script for ASP.NET Core 1.0, I generated a custom deployment script using:
npm install -g kuduscript
kuduscript -y --aspNetCore myproject.fsproj
Added the resulting deploy.cmd along with a .deployment:
[config]
command = deploy.cmd
A deployment triggered by a push to git works as expected now.