.Net Core With Docker

This example shows how to:

setup work environment for .NET Core,
create simple WebApi project with database usage,
setup database without needed to install anything more than docker,
setup Continuous Integration/Delivery pipeline to make sure that your code runns properly,
create test environment,
create prod environment.

Support

Feel free to create an issue if you have any questions or request for more explanation or samples. I also take Pull Requests!

If this repository helped you - I'd be more than happy if you join the group of my official supporters at:

Github Sponsors

Setup work environment for .NET Core

Install Docker
Install Visual Studio 2017, VisualStudioCode, Rider or your other favourite .NET IDE

Create new Web Api project

We will use default Visual Studio template for the WebApi project:

From the Visual Studio menu select: File => New => Project
Choose from the Templates => Visual C# => .Net Core an ASP.NET Core Web Application
Select also .NET Framework Version on the top (I suggest 4.7), select folder where you want to create new project and the choose the name.
In the new window select Web Api template. Unselect "Enable Docker Support" and leave Authentication settings as they are (so no authentication).
Open Package Manager Console and run dotnet restore command. It will get all needed NuGet packages for your application.
Now, when you have your basic app setup, you can click F5 to run it.
If everything went properly then you should see browser page with http://localhost:{port}/api/values and ["value1","value2"].

You can check the detailed changes in pull request

Add MSSQL Database to the Web Api project

Most of our applications needs to have the database. We'll use Entity Framework and MSSQL server in this example.

From the Package Manger Console run Install-Package Microsoft.EntityFrameworkCore.SqlServer and Install-Package Microsoft.EntityFrameworkCore.Tools. This will add Nuget Packages nessesary for the MSSQL server databasse usage.
Create Entity Class (eg. Task) and DbContext (eg. TasksDbContext).
You should get simmilar changes as in this commit.
Next step is to provide the connection string to the database. For this example we'll use LocalDB, which is distributed and installed automatically with the Visual Studio 2017 (if you're not using the Visual Studio then you can get it from this link).
We need to provide proper connection string to appsettings.json and pass it to the Entity Framework configuration in Startup.cs.
Having that configuration we can remove the dummy ValuesController and add new TasksController.
This example contains few important things derived from the best practices like:

async usage - it' much better for the performance perspective to use async/await in the api. What's crucial is that you cannot leave the async statement using database not awaited, because it may end up with your db connection not disposed properly,
returning proper Http Status Codes for the Api endpoints:
- OK - will return 200,
- NotFound- will return 404 result,
- Forbid - will return 403 status,
Good explanation of http status codes can be found here
few examples of the new usefull C# 6 syntax

You should also update your launchSettings.json to redirect you by default to the /tasks/ instead of the /values/ route.
If you run now you're application then you'll get following exception:

System.Data.SqlClient.SqlException: 'Cannot open database "NetCoreWithDocker" requested by the login. The login failed.

It basically means that it cannot login, because in fact there is no database that Entity Framework can login to. How to setup it? By using build in Migrations mechanism.
It's needed to setup the migrations (event the initial one), that will create database with object<=>relation mapping defined in the TasksDbContext. To do that open Package Manager Console and run: `Add-Migration InitialCreate](). This will automatically freeze the current model definiton in Current Model Snapshot and the define Initial Migration.
Now to create/update our database it's needed to run Update-Database from Package Manager Console.
You should now be able to run the application. If you did all of the steps properly then you should see browser page with http://localhost:{port}/api/values and []. That means that our TasksController returned empty list of tasks - because our database is currently empty.
Entity Framework also provides mechanism to add initial data (it's usefull for the eg. dictionaries or test data setup). Unfortunatelly it's needed to provide some boilerplate code for that. We need to do following steps:

12.1. Run Add-Migration InitialSeed in the Package Manager Console - this will generate empty migration.

12.2. Then we need to prepare our TasksDbContext to be also created not only from the depenedency injection. To do that we should create TasksDbContextFactory. This class will be used for the database configuration (eg. connection string) injection. It needs to implement IDbContextFactory<TasksDbContext> and unfortunatelly mimic some of the Startup Configuration reading functionality. Full implementation can be seen here.

12.3. Now we can use our factory class in our Initial Seed migration. In Up method we're defining our insertion, in Down method we're cleaning data (it' being run if something went wrong).

12.4. You need also to mark appsettings.Development.json as being coppied to ouput directory. To do that we need to add new settings in the NetCoreWithDocker.csproj.

Now you should be able to run Update-Database from Package Manager Console to apply seeding, start application by clicking F5 and see the results in the browser (eg. [{"id":1,"description":"Do what you have to do"},{"id":2,"description":"Really urgent task"},{"id":3,"description":"This one has really low priority"}])

You can check the detailed changes in pull request

Use MSSQL Database from Docker

Having docker being installed, now we can setup the docker container with MSSQL database (run on Linux). We'll use docker-compose tool, which simplyfies docker management, creation (especially for the multiple containers usage).
Let's add new folder docker in the root of our project. We will place there all of the docker configuration. Create also subfolder called mssql.
In the docker folder let's create docker-compose.yml - this will be our main configuration file. Docker configs are written in yaml syntax.
Our configuration:
```
version: "3"
services:
    mssql:
        image: "microsoft/mssql-server-linux"
        env_file:
            - mssql/variables.env
        ports:
            - "1433:1433"
```
It contains following sections:
- services - list of services (docker containers) that will be run,
- mssql - name of a service. It is provided by us, it could be named even xyz,
- env_file - reference to the files with environment needed for our service setup,
- ports - mapping of our port. This configuration mean that 1433 port from docker container will be mapped to our localhost 1433 port. Without that configuration port will be by default not accessible. It's also usefull if our local port is in use and we'd like to have different port assigned.
Now let's create variables.env file in the mssql folder and place there:
```
ACCEPT_EULA=Y
SA_PASSWORD=!QAZxsw2#EDC
```
- ACCEPT_EULA - is needed for accepting MSSQL Server licence terms,
- SA_PASSWORD - sa database user password
Having this setup ready we can open CMD from docker directory and run docker-compose up. This will download MSSQL server image from Docker Hub. It will also automatically start the server.
If everything went fine, then you should see SQL Server is now ready for client connections. in the CMD window.
Now we need to only update our connection strings in appsettings.json and appsettings.Development.json run Update-Database from Package Manager Console and we can run our application by clicking F5!
Piece and cake!
Summary of the docker-compose CLI can be found here. The most important commands are:

docker-compose up - as described above, gets images and starts containers,
docker-compose kill - kills running dockers,
docker-compose pull - pulls latest docker images,
docker system prune - clean up all containers that were get through pull and up commands. Useful for cleaning the disk space and making sure that you have the assumed version of docker (docker system prune + docker-compose up), or just resetting state of the docker container,
docker ps - lists all running docker containers.

You can check the detailed changes in pull request

Nuget packages to help you get started.

I gathered and generalized all of practices used in this tutorial/samples in Nuget Packages of maintained by me GoldenEye Framework. it provides set of base and bootstrap classes that helps you to reduce boilerplate code and help you focus on writing business code. See more in: