Table of Contents generated with DocToc

• Build a Microservice application with Spring Boot
  • Prerequisites
    • Setup local development environment
      • Docker Toolbox Notes
  • Cook your first service
    • REST API overview
    • Create domain models
    • Declare a Repository for Post entity
    • Create a domain service
    • Produces RESTful APIs
    • Exception Handling
    • Miscellaneous
  • Secures microservice
  • Run the application locally
  • Run all services via Docker Compose
  • Docker Swarm

Build a Microservice application with Spring Boot

Microservice is a very hot topic in these years, you can see it everywhere, there are a lots of books, blog entries, conference sessions, training courses etc are talking about it.

What is microservice?

Microservice is not a standard specification, so there is no official definition. Here I listed some well-known explanation from the communities.

Martin Fowler described it as the following in his article Microservice:

In short, the microservice architectural style is an approach to developing a single application as a suite of small services, each running in its own process and communicating with lightweight mechanisms, often an HTTP resource API. These services are built around business capabilities and independently deployable by fully automated deployment machinery. There is a bare minimum of centralized management of these services, which may be written in different programming languages and use different data storage technologies.

On the Wikipedia Microservice page, Microservice was defined as:

Microservices is a variant of the service-oriented architecture (SOA) architectural style that structures an application as a collection of loosely coupled services. In a microservices architecture, services should be fine-grained and the protocols should be lightweight. The benefit of decomposing an application into different smaller services is that it improves modularity and makes the application easier to understand, develop and test. It also parallelizes development by enabling small autonomous teams to develop, deploy and scale their respective services independently.[1] It also allows the architecture of an individual service to emerge through continuous refactoring. Microservices-based architectures enable continuous delivery and deployment.

Chris Richardson, the author of POJOs in Action and the creator of the original CloudFoundry.com, and also an advocator of Microservice, summarized Microservice as the following in the home page of Microservices.io.

Microservices - also known as the Microservice architecture - is an architectural style that structures an application as a collection of loosely coupled services, which implement business capabilities. The Microservice architecture enables the continuous delivery/deployment of large, complex applications. It also enables an organization to evolve its technology stack.

There are some common characteristics can be used to describe a Microservice based application.

• A Microservice application should be consisted of a collection of small services. One service is not Microservice. Every service is fine-grained, and target to perform a small function. So Microservice was described as fine-grained SOA or SOA done right in some articles. So This is the main difference from traditional monolithic applications.

• Every service should have its own independent life cycle. Every service can be developed and deployed independently, if you are using a CI/CD automation service, every service should be done in a complete DevOps pipeline flow, but not affect others.

• Service-to-service communication is based on light-weight protocols, eg. HTTP based REST APIs for synchronous communication, WebSocket for asynchronous messages, MQTT/AMQP protocol for varied messaging from client or devices(eg. IOT applications).

• The organization or team structures should be changed simultaneously when you are embracing Microservice architecture. You have to break your traditional organization tree. In traditional application development, your teams are organized by roles, eg architects, database administrators, developers, testers, operators etc. In the development stage of a Microservice based application, a team should be responsible for the whole DevOps lifecycle of one or more services.

Microservice componentizes your application into small services(componentized applications), and make it more maintainable and scalable. In this demo application, I will show you building a Microservice application via Spring Boot.

Migrate to Microservice architecture

Contrast with Microservice applications, traditional layered enterprise applications were called monolithic applications.

In the past years, I have created some samples to demonstrate different technology stack, such as REST APIs sample with Spring MVC, REST APIs sample with Spring Boot, in these samples, the backends are monolithic applications and they are based on the same model prototype, a blog application.

• A user can log in with an existed account, or sign up a new account.
• An authenticated user can create a new post.
• An authenticated user can update his/her posts.
• An authenticated user who has ADMIN role can delete a post directly.
• All users(who are authenticated or anonymous) can view posts.
• An authenticated user can add comments to an existed post.
• ...

No doubt these monolithic backend applications are easy to develop and deploy, but as time goes by, when the application becomes more complex, the backend will be problematic, you maybe face some barriers which block you to the next stages.

• When you apply a change, you have to redeploy the whole backend application even it is just a small fix. The application may be stopped to work for some minutes or some hours.
• When you scale your applications and deploy multi copies of the backend applications behinds a load balance server, the transactional consistence will be a new challenge.
• The database itself will be a huge performance bottleneck when the concurrency of incoming requests are increasing.

Microservice architecture addresses these problems, including:

1. Smaller services are easier to develop and deploy, when you upgrade one of the services, you do not need to shut down all services in production.
2. ACID can not satisfy the scenario of those long run workflow which across several services, although it is still a good option within a single service, but for these long run transactions, a stateful Saga or workflow solution fills this field.
3. A service can has its own database, and only responsible for storing data of this service itself. Traditional complex queries will become a big challenge, in Microservice architecture, it could need to query multi independent database and aggregate the query results. CQRS, Event Store can save these. Perform commands in standalone services, and execute queries in another service which has marshal view of the data and was synced with messaging from events triggered by other services.

Follow the Bounded Context concept of DDD(Domain Driven Design), we break the backend monolithic application into three small services, including:

• An auth-service is serving the operations of signin, signup and signout.
• A user-service is responsible for user management.
• A post-service exposes APIs for a simple CMS, including posts and comments.
• An API Gateway which is just responsible for routing the incoming requests to downstream services.
• The databases are also aligned to Microservice architecture, and user-service and post-service have their own databases, a Redis is used for sharing session between services, and to simplify the security.

As mentioned, if there is a legacy application planned to migrate to Microservice architecture, you can follow the following steps to extract some domain into a standalone service.

1. Find the domains which are easiest to separate from the main application, eg, posts and comments in our application.
2. Use an identifier object in the entity links instead of the hard relations of entities outside of this domain. eg. use a Username which stands for a unique username of a User entity, and erase the direct connection to User entity.
3. Move the related data to a standalone database, and connect to this new database in your service.

When I start a new project, should I embrace Microservice architecture right now?

Although we are talking about Microservice in this post, I still suggest you start building your application in a monolithic architecture if you know little about the complexity of Microservice, it could be consisted of a RESTful backend and an SPA based frontend UI. In the initial development stage, either monolithic architecture or Microservice, you have to spend lots of time on clarifying the problem domains, defining the bounded context etc. Starting a monolithic application is still valuable when you are ready for migrating to Microservice architecture.

Cook our first service

This sample application is built on the newest Spring technology stack, including Spring Boot, Spring Data, Spring Security, etc.

• Every small service is a Spring Boot application. Every service will be packaged as a jar file and use the embedded Tomcat as target runtime to serve the services.
• Every small service owns its database, eg. we use MySQL as the backing database for auth-service, and PostgreSQL for the post-service.
• Spring Data is used for simplifying data operations.
• Spring Session provides a simple strategy to generate and validate header based authentication token via sharing sessions in a backing session repository, in this sample we use Redis as session storage.
• Spring Security is responsible for protecting RESTful APIs.

Follow the 12 factors application guide, I suggest you use Docker in both development and production environment to make sure the same code base works well in different environments.

In this section, we will build our first service, post-service, which is designated to exposes REST APIs to clients.

Prerequisites

I assume you have some experience of Spring, and know well about the REST convention, esp the CHAPTER 5: Representational State Transfer (REST) from Roy Fielding's dissertation: Architectural Styles and the Design of Network-based Software Architectures.

And you have also installed the following software.

• Oracle Java 8 SDK
• Apache Maven
• Gradle if you prefer Gradle as build tools
• Your favorite IDE, including :
  • NetBeans IDE
  • Eclipse IDE (or Eclipse based IDE, Spring ToolSuite is highly recommended)
  • Intellij IDEA

Setup local development environment

Make sure you have installed the latest Docker, Docker Compose and Docker Machine, more info please refer to the installation guide from Docker official website.

NOTE: I used the official Docker Toolbox in development stage under Windows 10, you can also use Docker for Windows instead.

Docker Compose allow you start up the dependent infrastructural services(such as Database etc) via a single docker-compose command.

docker-compose up

We will use MySQL, PostgreSQL and Redis in this demo, the following is a sample docker-compose.yml file.

version: '3.3' # specify docker-compose version

services:    
  userdb:
    container_name: userdb
    image: mysql
    ports:
      - "3306:3306"
    environment:
      MYSQL_ROOT_PASSWORD: mysecret
      MYSQL_USER: user
      MYSQL_PASSWORD: password
      MYSQL_DATABASE: userdb
    volumes:
      - ./data/userdb:/var/lib/mysql
      
  postdb:
    container_name:  postdb
    image: postgres
    ports:
      - "5432:5432"
    restart: always
    environment:
      POSTGRES_PASSWORD: password
      POSTGRES_DB: postdb
    volumes:
      - ./data/postdb:/var/lib/postgresql     
      

  redis:
    container_name: redis
    image: redis
    ports:
      - "6379:6379"

Docker Toolbox Notes

If you are using Docker Toolbox, create a new machine for this project.

$ docker-machine create -d virtualbox --engine-registry-mirror https://docker.mirrors.ustc.edu.cn springms

NOTE: The --engine-registry-mirror https://docker.mirrors.ustc.edu.cn will add a docker registry mirror setting in docker-machine specific config.json. For most of Chinese users, using a local mirror will speed up the Docker images downloading.

Then switch to the new created machine springms, and set the environment variables.

eval "$(docker-machine env springms)"

Forward the virtualbox ports to your local system, thus you can access the servers via localhost instead of the docker machine IP address.

 VBoxManage modifyvm "springms" --natpf1 "tcp-port3306,tcp,,3306,,3306"
 VBoxManage modifyvm "springms" --natpf1 "tcp-port3307,tcp,,3307,,3307"
 VBoxManage modifyvm "springms" --natpf1 "tcp-port5672,tcp,,5672,,5672"
 VBoxManage modifyvm "springms" --natpf1 "tcp-port15672,tcp,,15672,,15672"
 VBoxManage modifyvm "springms" --natpf1 "tcp-port6379,tcp,,6379,,6379"
 VBoxManage modifyvm "springms" --natpf1 "tcp-port27017,tcp,,27017,,27017"

Then run the dependent servers via docker-compose command line.

Generate project skeleton

With Spring Initializr, you can get a Spring Boot based project skeleton in seconds.

Open your browser, go to Spring Initializr page, fill the following essential fields for a project.

1. Choose Java as programming language.
2. Select the latest version of Spring Boot, 2.0.0.M7 is the latest milestone at the moment when I wrote this post.
3. Search and select the required facilities will be used in your project, such as Web, Data JPA, Data Redis, Security, Session, Lombok etc.
4. Set project name(maven artifact id) to post-service.

Click Generate Project button or press ALT+ENTER keys to generate the project skeleton for downloading in your browser.

After downloading the generated archive, extract the files into your local disk and import it into your favorite IDE.

REST API overview

Following the REST convention and HTTP protocol specification, the REST APIs of post-service are designed as the following table.

Create an entity

In DDD, an entity is a persistent object in DDD concept, JPA @Entity a is a good match.

Create our first entity Post.

@Data
@Builder
@NoArgsConstructor
@AllArgsConstructor
@Entity
class Post extends AuditableEntity {

    @JsonView(View.Summary.class)
    @NotEmpty
    private String title;
    
    @NotEmpty
    private String slug;

    @JsonView(View.Public.class)
    @NotEmpty
    private String content;

    @Enumerated(EnumType.STRING)
    @Builder.Default
    @JsonView(View.Summary.class)
    private Status status = Status.DRAFT;

    static enum Status {
        DRAFT,
        PUBLISHED
    }
      
    @PrePersist
    public void slugify(){
        this.slug = new Slugify().slugify(this.title);
    }

}

@Data, @Builder, @NoArgsConstructor and @AllArgsConstructor are from project Lombok, which provides some helper annotations to make your source codes clean. With @Data, you can remove the tedious setters, getters of all fields, and the generic equals, hashCode, toString methods. @Builder will generate an inner builder class. @NoArgsConstructor will create a none-argument constructor, @AllArgsConstructor will take all fields as constructor arguments.

These annotations will be handled by JDK Annotation Processing Tooling, and generate code fragment in class files at compile time.

@Entity indicates Post is a standard JPA Entity.

@PrePersist is a JPA lifecycle hook. The @PrePersist annotated methods will be executed before the entity is persisted. We use post slug as the unique identifier of a Post, and we use slugify() method to generate the post slug automatically.

AuditableEntity is a helper class to centralize some common fields of a JPA entity in one place.

@Data
@MappedSuperclass
@EntityListeners(value = AuditingEntityListener.class)
public abstract class AuditableEntity extends PersistableEntity {

    public AuditableEntity() {
    }

    @CreatedDate
    @JsonView(View.Summary.class)
    protected LocalDateTime createdDate;

    @Embedded
    @AttributeOverrides(value = {
        @AttributeOverride(name = "username", column = @Column(name = "author"))
    })
    @CreatedBy
    @JsonView(View.Summary.class)
    protected Username author;

}

@CreatedDate and @CreatedBy will fill in the creation date timestamp and the current user if the data auditing feature is enabled.

Use a standalone @Configuration bean to configure Spring Data JPA auditing.

@Configuration
@EnableJpaAuditing(auditorAwareRef = "auditorAware")
@Slf4j
public class DataJpaConfig {

    @Bean
    @Scope(value = ConfigurableBeanFactory.SCOPE_PROTOTYPE)
    public AuditorAware<Username> auditorAware() {

        Authentication authentication = SecurityContextHolder.getContext().getAuthentication();

        log.debug("current authentication:" + authentication);

        if (authentication == null || !authentication.isAuthenticated()) {
            return () -> Optional.<Username>empty();
        }

        return () -> Optional.of(
            Username.builder()
                .username(((UserDetails) authentication.getPrincipal()).getUsername())
                .build()
        );

    }
}

AuditorAware bean is required when you want to set auditor automatically. The population work is done by JPA @EntityListener, note there is a @EntityListeners(value = AuditingEntityListener.class) already added on the AuditableEntity class.

Have a look at the base PersistableEntity, it just defines the identity field of a JPA entity.

@Data
@MappedSuperclass
public abstract class PersistableEntity implements Serializable {
    
    @Id
    @GeneratedValue(strategy = GenerationType.AUTO)
    @JsonView(View.Summary.class)
    protected Long id;

    public PersistableEntity() {
    }
    
}

Similarly, create an another entity Comment.

@Data
@Builder
@NoArgsConstructor
@AllArgsConstructor
@Entity
public class Comment extends AuditableEntity {


    @NotEmpty
    @Size(min = 10)
    private String content;


    @Embedded
    @AttributeOverrides(
        value = {
            @AttributeOverride(name = "slug", column = @Column(name = "post_slug"))
        }
    )
    @JsonIgnore
    private Slug post;


}

NOTE: we do not user a JPA @OneToMany or @ManyToOne to connect two entities, but use a simple Post Slug identifier object instead. If one day this service becomes heavy, we could split comments into another standalone service.

Declare a Repository

In DDD, a Repository is responsible for retrieving entities from or saving back to a Repository. Spring Data Repository interface and Spring Data JPA specific JpaRepository interface are a good match with Repository concept in DDD.

public interface PostRepository extends JpaRepository<Post, Long>, JpaSpecificationExecutor<Post> {

    Optional<Post> findBySlug(String slug);
    
}

Create a domain service

@Service
@Transactional
public class PostService {

    @Inject
    private PostRepository postRepository;
    

    public Post createPost(PostForm form) {
        Post _post = Post.builder()
            .title(form.getTitle())
            .content(form.getContent())
            .build();
        
        Post saved = this.postRepository.save(_post);
        
        return saved;
    }

    public Post updatePost(String slug, PostForm form) {
        Post _post = this.postRepository.findBySlug(slug).orElseThrow(
            ()-> {
                return new PostNotFoundException(slug);
            }
        );
        
        _post.setTitle(form.getTitle());
        _post.setContent(form.getContent());
        
       Post saved =  this.postRepository.save(_post);
       
       return saved;
    }

    public void deletePost(String slug) {
        this.postRepository.delete(this.postRepository.findBySlug(slug).orElseThrow(
            () -> {
                return new PostNotFoundException(slug);
            }
        ));
    }

}

In the PostService, the main purpose is treating with exceptions when creating or update a post. In a real world application, you could handle domain events, eg. Post is published etc.

Produces RESTful APIs

Let's expose RESTful APIs for Post via PostController.

@RestController
@RequestMapping("/posts")
@Slf4j
public class PostController {

    private PostService postService;

    private PostRepository postRepository;

    private CommentRepository commentRepository;

    public PostController(PostService postService, PostRepository postRepository, CommentRepository commentRepository) {
        this.postService = postService;
        this.postRepository = postRepository;
        this.commentRepository = commentRepository;
    }

    @GetMapping()
    @JsonView(View.Summary.class)
    public ResponseEntity<Page<Post>> getAllPosts(
        @RequestParam(value = "q", required = false) String keyword, //
        @RequestParam(value = "status", required = false) Post.Status status, //
        @PageableDefault(page = 0, size = 10, sort = "createdDate", direction = Direction.DESC) Pageable page) {

        log.debug("get all posts of [email protected]" + keyword + ", status @" + status + ", [email protected]" + page);

        Page<Post> posts = this.postRepository.findAll(PostSpecifications.filterByKeywordAndStatus(keyword, status), page);

        return ok(posts);
    }

    @GetMapping(value = "/{slug}")
    @JsonView(View.Public.class)
    public ResponseEntity<Post> getPost(@PathVariable("slug") String slug) {

        log.debug("get postsinfo by slug @" + slug);

        Post post = this.postRepository.findBySlug(slug).orElseThrow(
            () -> {
                return new PostNotFoundException(slug);
            }
        );

        log.debug("get post @" + post);

        return ok(post);
    }

    @PostMapping()
    public ResponseEntity<Void> createPost(@RequestBody @Valid PostForm post, HttpServletRequest request) {

        log.debug("create a new [email protected]" + post);

        Post saved = this.postService.createPost(post);

        log.debug("saved post id is @" + saved.getId());
        URI createdUri = ServletUriComponentsBuilder
            .fromContextPath(request)
            .path("/posts/{slug}")
            .buildAndExpand(saved.getSlug())
            .toUri();

        return created(createdUri).build();
    }

    @PutMapping(value = "/{slug}")
    public ResponseEntity<Void> updatePost(@PathVariable("slug") String slug, @RequestBody @Valid PostForm form) {

        log.debug("update post by id @" + slug + ", form [email protected]" + form);

        this.postService.updatePost(slug, form);

        return noContent().build();
    }

    @DeleteMapping(value = "/{slug}")
    public ResponseEntity<Void> deletePostById(@PathVariable("slug") String slug) {

        log.debug("delete post by id @" + slug);

        this.postService.deletePost(slug);

        return noContent().build();
    }

    @GetMapping(value = "/{slug}/comments")
    public ResponseEntity<Page<Comment>> getCommentsOfPost(
        @PathVariable("slug") String slug,
        @PageableDefault(page = 0, size = 10, sort = "createdDate", direction = Direction.DESC) Pageable page) {

        log.debug("get comments of [email protected]" + slug + ", [email protected]" + page);

        Page<Comment> commentsOfPost = this.commentRepository.findByPost(new Slug(slug), page);

        log.debug("get post comment size @" + commentsOfPost.getTotalElements());

        return ok(commentsOfPost);
    }

    @PostMapping(value = "/{slug}/comments")
    public ResponseEntity<Void> createComment(
        @PathVariable("slug") @NotNull String slug, @RequestBody CommentForm comment, HttpServletRequest request) {

        log.debug("new comment of [email protected]" + slug + ", comment" + comment);

        Comment _comment = Comment.builder()
            .post(new Slug(slug))
            .content(comment.getContent())
            .build();

        Comment saved = this.commentRepository.save(_comment);

        log.debug("saved comment @" + saved.getId());

        URI location = ServletUriComponentsBuilder
            .fromContextPath(request)
            .path("/posts/{slug}/comments/{id}")
            .buildAndExpand(slug, saved.getId())
            .toUri();

         return created(location).build();
    }

}

getAllPosts method accepts a q (keyword) and a status (post status) and a Pageable as query parameters, it returns a Page<Post> result. The postRepository.findAll method accepts a Specification object. Specification is a wrapper class of JPA 2.0 criteria APIs, which provides effective type safe query condition building.

public class PostSpecifications {

    private PostSpecifications() {
    }

    public static Specification<Post> filterByKeywordAndStatus(
        final String keyword,//
        final Post.Status status) {
        return (Root<Post> root, CriteriaQuery<?> query, CriteriaBuilder cb) -> {
            List<Predicate> predicates = new ArrayList<>();
            if (StringUtils.hasText(keyword)) {
                predicates.add(
                    cb.or(
                        cb.like(root.get(Post_.title), "%" + keyword + "%"),
                        cb.like(root.get(Post_.content), "%" + keyword + "%")
                    )
                );
            }

            if (status != null) {
                predicates.add(cb.equal(root.get(Post_.status), status));
            }

            return cb.and(predicates.toArray(new Predicate[predicates.size()]));
        };
    }

}

According to the REST convention and HTTP protocol, a HTTP POST Method is used to create a new resource, it can return a 201 HTTP status code with the new created resource URI as HTTP header Location. And for update and delete operations on resource, return a 204 HTTP status. In the above codes, we apply these simple rules.

Exception Handling

As mentioned above, in our PostService, I have added some extra steps to check the existence of a post by id in the updatePost and deletePost methods. If it is not found throw a PostNotFoundException.

public class PostNotFoundException extends RuntimeException {

    private String slug;

    public PostNotFoundException(String slug) {
        super("post:" + slug + " was not found");
        this.slug = slug;
    }

    public String getSlug() {
        return slug;
    }
    
}

And we will handle this exception in a common class annotated with @RestControllerAdvice. When a PostNotFoundException is caught, notFound method will handle it convert the exception to a friendly message body and return a HTTP 404 status code to the client.

@RestControllerAdvice
public class PostExceptionHandler {

    @ExceptionHandler(PostNotFoundException.class)
    public ResponseEntity notFound(PostNotFoundException ex, WebRequest req) {
        Map<String, String> errors = new HashMap<>();
        errors.put("entity", "POST");
        errors.put("id", "" + ex.getSlug());
        errors.put("code", "not_found");
        errors.put("message", ex.getMessage());

        return ResponseEntity.status(HttpStatus.NOT_FOUND).body(errors);
    }

}

Miscellaneous

In a real world application, when you fetch post list, you maybe do not want to show all fields of the post. It is easy to control the representation view sent to client by customizing Jackson JsonView.

public final class View {

    interface Summary {
    }

    interface Public extends Summary {
    }
}

In the Post class, add the following annotations to its fields.

class Post extends AuditableEntity {

    @JsonView(View.Summary.class)
    private String title;
    

    @JsonView(View.Public.class)
    private String content;

    @JsonView(View.Summary.class)
    private Status status = Status.DRAFT;

}

In the PostController, add a @JsonView annotation.

@JsonView(View.Summary.class)
public ResponseEntity<Page<Post>> getAllPosts()

Thus only the Summary labeled fields will be included in the result of getAllPosts.

Another small issue you could have found is the Page object serialized result looks a little tedious, too much unused fields from Pageable are included in the json result.

{
  "content" : [ {
    "title" : "test post 2",
    "slug" : "test-post-2",
    "status" : "DRAFT",
    "id" : 2,
    "createdDate" : "2017-05-25T06:53:30",
    "author" : {
      "username" : "user"
    }
  }, {
    "title" : "test post",
    "slug" : "test-post",
    "status" : "DRAFT",
    "id" : 1,
    "createdDate" : "2017-05-25T06:52:45",
    "author" : {
      "username" : "user"
    }
  } ],
  "pageable" : {
    "sort" : {
      "sorted" : true,
      "unsorted" : false
    },
    "pageSize" : 10,
    "pageNumber" : 0,
    "offset" : 0,
    "paged" : true,
    "unpaged" : false
  },
  "last" : true,
  "totalElements" : 2,
  "totalPages" : 1,
  "sort" : {
    "sorted" : true,
    "unsorted" : false
  },
  "numberOfElements" : 2,
  "first" : true,
  "size" : 10,
  "number" : 0
}

Create a @JsonComponent bean to customize the serialized json result.

@JsonComponent
public class PageJsonSerializer extends JsonSerializer<PageImpl> {

    @Override
    public void serialize(PageImpl value, JsonGenerator gen, SerializerProvider serializers) throws IOException, JsonProcessingException {
        gen.writeStartObject();
        gen.writeNumberField("number", value.getNumber());
        gen.writeNumberField("numberOfElements", value.getNumberOfElements());
        gen.writeNumberField("totalElements", value.getTotalElements());
        gen.writeNumberField("totalPages", value.getTotalPages());
        gen.writeNumberField("size", value.getSize());
        gen.writeFieldName("content");
        serializers.defaultSerializeValue(value.getContent(), gen);
        gen.writeEndObject();
    }

}

When this bean is activated, the result cloud look like the following:

{
  "content" : [ {
    "title" : "test post 2",
    "slug" : "test-post-2",
    "status" : "DRAFT",
    "id" : 2,
    "createdDate" : "2017-05-25T06:53:30",
    "author" : {
      "username" : "user"
    }
  }, {
    "title" : "test post",
    "slug" : "test-post",
    "status" : "DRAFT",
    "id" : 1,
    "createdDate" : "2017-05-25T06:52:45",
    "author" : {
      "username" : "user"
    }
  } ],
  "numberOfElements" : 2,
  "totalElements" : 2,
  "totalPages" : 1,
  "size" : 10,
  "number" : 0
}

The details of auth-service and user-service, please check the source codes and explore them yourself.

Secures Microservice

Let's have a look at how a user get authentication in this demo.

1. A user try to get authentication from auth-service using usename and password.
2. If it is a valid user and it is authenticated successfully, the response header will include a X-AUTH-TOKEN header.
3. Extract the value of X-AUTH-TOKEN header, and add X-AUTH-TOKEN header into the new request to get access permission of the protected resource, such as APIs in post-service.

We use Spring Session and Redis to archive this purpose.

In all services, we add the following codes to resolve Session by HTTP header instead of Cookie.

@Configuration
public class RedisSessionConfig {

    @Bean
    public HttpSessionIdResolver httpSessionStrategy() {
        return HeaderHttpSessionIdResolver.xAuthToken();
    }

}

And add the follow configuration in the application.yml to tell Spring to use Redis as session store.

spring:
  session: 
    store-type: redis

In auth-service, use a controller to serve user authentication.

@RequestMapping(value = "/auth")
@RestController
public class AuthenticationController {

	@PostMapping(value = "/signin")
    public AuthenticationResult signin(
        @RequestBody @Valid AuthenticationRequest authenticationRequest,
        HttpServletRequest request) {
        
        if (log.isDebugEnabled()) {
            log.debug("signin form  [email protected]" + authenticationRequest);
        }
        
        return this.handleAuthentication(
            authenticationRequest.getUsername(),
            authenticationRequest.getPassword(),
            request);
    }
    
    private AuthenticationResult handleAuthentication(
        String username,
        String password,
        HttpServletRequest request) {
        
        final UsernamePasswordAuthenticationToken token = new UsernamePasswordAuthenticationToken(
            username,
            password
        );
        
        final Authentication authentication = this.authenticationManager
            .authenticate(token);
        
        SecurityContextHolder.getContext().setAuthentication(authentication);
        
        final HttpSession session = request.getSession(true);
        
        session.setAttribute(
            HttpSessionSecurityContextRepository.SPRING_SECURITY_CONTEXT_KEY,
            SecurityContextHolder.getContext());
        
        return AuthenticationResult.builder()
            .name(authentication.getName())
            .roles(authentication.getAuthorities().stream().map(r -> r.getAuthority()).collect(Collectors.toList()))
            .token(session.getId())
            .build();
    }
	
	...

When you are authenticated, the /auth/signin endpoint will return userinfo and token(session id) in the result.

To protect the resource APIs, just add a SecurityConfig. The following is a configuration for post-service. All GET methods are permitted, and when DELETE a post, you should have a ADMIN role.

@Configuration
@Slf4j
public class SecurityConfig {
    
    @Bean
    public WebSecurityConfigurerAdapter securityConfigBean(){
        
        return new  WebSecurityConfigurerAdapter() {

            @Override
            protected void configure(HttpSecurity http) throws Exception {
                // We need this to prevent the browser from popping up a dialog on a 401
                http
                    .httpBasic()
                    .and()
                        .authorizeRequests()
                        .antMatchers(HttpMethod.GET, "/posts/**").permitAll()
                        .antMatchers(HttpMethod.DELETE, "/posts/**").hasRole("ADMIN")
                        .anyRequest().authenticated()
                    .and()
                        .csrf().disable();
            }
        };   
    }
}

Let's try to run the demo in local system.

Run the application locally

Make sure the dependent servers are running by executing docker-compose up.

Enter the root folder of every service, execute the following command to start up them one by one.

mvn spring-boot:run // run in user-service, auth-service, post-service

The following endpoints will be provided.

Follow the authentication flow to have a try.

When all service are running successfully, firstly try to get authentication.

curl -v  http://localhost:8000/user -u user:test123
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
* Connected to localhost (::1) port 8000 (#0)
* Server auth using Basic with user 'user'
> GET /user HTTP/1.1
> Host: localhost:8000
> Authorization: Basic dXNlcjp0ZXN0MTIz
> User-Agent: curl/7.54.0
> Accept: */*
>
< HTTP/1.1 200
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
< X-Auth-Token: 49090ba7-e641-45e3-935b-894a43b85f62
< Content-Type: application/json;charset=UTF-8
< Transfer-Encoding: chunked
< Date: Mon, 15 May 2017 09:29:14 GMT
<
{"name":"user","roles":["USER"]}* Connection #0 to host localhost left intact

You will see a X-Auth-Token header in the response.

Put this header into a new request when you want to access the protected resources in another resource server.

curl -v  http://localhost:8001/user -H "x-auth-token: 49090ba7-e641-45e3-935b-894a43b85f62"

Try to add some posts data:

>curl -v  http://localhost:8002/posts 
-H "x-auth-token:  49090ba7-e641-45e3-935b-894a43b85f62" 
-H "Accept: application/json" 
-H "Content-Type: application/json;charset=UTF-8" 
-X POST 
-d "{\"title\": \"test post\", \"content\":\"test content of post\"}"

You will see the result. It returns 201 status, and set Location header to the new created Post.

Note: Unnecessary use of -X or --request, POST is already inferred.
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
* Connected to localhost (::1) port 8002 (#0)
> POST /posts HTTP/1.1
> Host: localhost:8002
> User-Agent: curl/7.54.0
> x-auth-token:  49090ba7-e641-45e3-935b-894a43b85f62
> Accept: application/json
> Content-Type: application/json;charset=UTF-8
> Content-Length: 56
>
* upload completely sent off: 56 out of 56 bytes
< HTTP/1.1 201
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
< Location: http://localhost:8002/posts/4
< Content-Length: 0
< Date: Thu, 18 May 2017 06:54:40 GMT

Fetch the new created post.

curl -v  http://localhost:8002/posts/4 -H "Accept: application/json"
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
* Connected to localhost (::1) port 8002 (#0)
> GET /posts/4 HTTP/1.1
> Host: localhost:8002
> User-Agent: curl/7.54.0
> Accept: application/json
>
< HTTP/1.1 200
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
< Content-Type: application/json;charset=UTF-8
< Transfer-Encoding: chunked
< Date: Thu, 18 May 2017 06:59:42 GMT
<
{"id":4,"title":"test post","content":"test content of post","status":"DRAFT","author":null,"createdDate":null}*

Run all services via Docker Compose

Prepare a Dockfile for every service.

FROM frolvlad/alpine-oraclejdk8:slim
VOLUME /tmp
ADD ./target/post-service-0.0.1-SNAPSHOT.jar app.jar
RUN sh -c 'touch /app.jar'
ENV JAVA_OPTS=""
ENTRYPOINT [ "sh", "-c", "java $JAVA_OPTS -Djava.security.egd=file:/dev/./urandom -jar /app.jar" ]

The Dockerfile in auth-service and user-service are similar, just replaced the maven build target file.

FROM frolvlad/alpine-oraclejdk8:slim
VOLUME /tmp
ADD ./target/auth-service-0.0.1-SNAPSHOT.jar app.jar
RUN sh -c 'touch /app.jar'
ENV JAVA_OPTS=""
ENTRYPOINT [ "sh", "-c", "java $JAVA_OPTS -Djava.security.egd=file:/dev/./urandom -jar /app.jar" ]

Create a Dockerfile for ngnix.

# Set nginx base image
FROM nginx 

#RUN mkdir /etc/nginx/ssl  
#COPY ssl /etc/nginx/ssl 

# Copy custom configuration file from the current directory
COPY nginx.conf /etc/nginx/nginx.conf
  
#COPY www /usr/share/nginx/www  
#COPY archive /usr/share/nginx/archive

And the content of ngnix.conf.

worker_processes 1;

events { worker_connections 1024; }

http {
    sendfile on;

	server {
		listen 80;
		server_name localhost;

		proxy_set_header Host $host;
		proxy_set_header X-Forwarded-For $remote_addr;


		location /users {
			proxy_pass http://user-service:8001;
		}
		location /posts {
			proxy_pass http://post-service:8002;
		}
		location / {
			proxy_pass http://auth-service:8000;
		}
	}
}

Create a standalone docker-compose.local.yml file to run all services.

version: '3.1' # specify docker-compose version

services:

  nginx-proxy:
    image: hantsy/nginx-proxy
    container_name: nginx-proxy
    build: 
      context: ./nginx
      dockerfile: Dockerfile
    depends_on:
      - auth-service
      - user-service
      - post-service
    ports:
      - "80:80"
      
  auth-service:
    image: hantsy/auth-service
    container_name: auth-service
    build: 
      context: ./auth-service # specify the directory of the Dockerfile
      dockerfile: Dockerfile
    environment:
      USERDB_URL: jdbc:mysql://userdb:3306/userdb
      REDIS_HOST: redis
    ports:
      - "8000:8000" #specify ports forewarding
    depends_on:
      - userdb
      - redis
      
  user-service: 
    image: hantsy/user-service
    container_name: user-service
    build: 
      context: ./user-service
      dockerfile: Dockerfile
    environment:
      USERDB_URL: jdbc:mysql://userdb:3306/userdb
      REDIS_HOST: redis
    ports:
      - "8001:8001" #specify ports forewarding
    depends_on:
      - userdb
      - redis
  
  post-service: 
    image: hantsy/post-service
    container_name: post-service
    build: 
      context: ./post-service
      dockerfile: Dockerfile
    environment:
      POSTDB_URL: jdbc:mysql://postdb:3306/postdb
      REDIS_HOST: redis
    ports:
      - "8002:8002" #specify ports forewarding
    depends_on:
      - postdb
      - redis 

Run all services in your local system or a staging server.

Build the project via mvn command.

mvn clean package -DskipTests

Then run the following command to run all services.

docker-compose -f docker-compose.yml -f docker-compose.local.yml up --build

The --build parameter tells Docker build Docker images for all services firstly, then create containers based on the built images.

We have run a Nginx a reverse proxy, all APIs can be accessed through a single entry.

The following services will be provided.

Get authentication.

curl -v  http://localhost/user -u user:test123

>
< HTTP/1.1 200
< Server: nginx/1.13.0
< Date: Thu, 25 May 2017 06:49:52 GMT
< Content-Type: application/json;charset=UTF-8
< Transfer-Encoding: chunked
< Connection: keep-alive
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
< X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8
<
{"name":"user","roles":["USER"]}* Connection #0 to host localhost left intact

Create a new post.

curl -v  http://localhost/posts -X POST -H "X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8" -H "Content-Type:application/json" -d "{\"title\": \"test post\", \"content\":\"test content of post\"}"
Note: Unnecessary use of -X or --request, POST is already inferred.
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
*   Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to localhost (127.0.0.1) port 80 (#0)
> POST /posts HTTP/1.1
> Host: localhost
> User-Agent: curl/7.54.0
> Accept: */*
> X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8
> Content-Type:application/json
> Content-Length: 56
>
* upload completely sent off: 56 out of 56 bytes
< HTTP/1.1 201
< Server: nginx/1.13.0
< Date: Thu, 25 May 2017 06:52:46 GMT
< Content-Length: 0
< Connection: keep-alive
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
< Location: http://localhost/posts/1
<
* Connection #0 to host localhost left intact

Create another new post.

curl -v  http://localhost/posts -X POST -H "X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8" -H "Content-Type:application/json" -d "{\"title\": \"test post 2\", \"content\":\"test content of post 2\"}"
Note: Unnecessary use of -X or --request, POST is already inferred.
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
*   Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to localhost (127.0.0.1) port 80 (#0)
> POST /posts HTTP/1.1
> Host: localhost
> User-Agent: curl/7.54.0
> Accept: */*
> X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8
> Content-Type:application/json
> Content-Length: 60
>
* upload completely sent off: 60 out of 60 bytes
< HTTP/1.1 201
< Server: nginx/1.13.0
< Date: Thu, 25 May 2017 06:53:29 GMT
< Content-Length: 0
< Connection: keep-alive
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
< Location: http://localhost/posts/test-post-2
<
* Connection #0 to host localhost left intact

Verify the created posts.

curl -v  http://localhost/posts  -H "Accpet:application/json"
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
*   Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to localhost (127.0.0.1) port 80 (#0)
> GET /posts HTTP/1.1
> Host: localhost
> User-Agent: curl/7.54.0
> Accept: */*
> Accpet:application/json
>
< HTTP/1.1 200
< Server: nginx/1.13.0
< Date: Thu, 25 May 2017 06:53:58 GMT
< Content-Type: application/json;charset=UTF-8
< Transfer-Encoding: chunked
< Connection: keep-alive
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
<
{
  "content" : [ {
    "title" : "test post 2",
    "slug" : "test-post-2",
    "status" : "DRAFT",
    "id" : 2,
    "createdDate" : "2017-05-25T06:53:30",
    "author" : {
      "username" : "user"
    }
  }, {
    "title" : "test post",
    "slug" : "test-post",
    "status" : "DRAFT",
    "id" : 1,
    "createdDate" : "2017-05-25T06:52:45",
    "author" : {
      "username" : "user"
    }
  } ],
  "pageable" : {
    "sort" : {
      "sorted" : true,
      "unsorted" : false
    },
    "pageSize" : 10,
    "pageNumber" : 0,
    "offset" : 0,
    "paged" : true,
    "unpaged" : false
  },
  "last" : true,
  "totalElements" : 2,
  "totalPages" : 1,
  "sort" : {
    "sorted" : true,
    "unsorted" : false
  },
  "numberOfElements" : 2,
  "first" : true,
  "size" : 10,
  "number" : 0
}* Connection #0 to host localhost left intact

Create a comment for "test post 2".

curl -v  http://localhost/posts/test-post-2/comments -X POST -H "X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8" -H "Content-Type:application/json" -d "{ \"content\":\"conmment content of post 2\"}"
Note: Unnecessary use of -X or --request, POST is already inferred.
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
*   Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to localhost (127.0.0.1) port 80 (#0)
> POST /posts/test-post-2/comments HTTP/1.1
> Host: localhost
> User-Agent: curl/7.54.0
> Accept: */*
> X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8
> Content-Type:application/json
> Content-Length: 41
>
* upload completely sent off: 41 out of 41 bytes
< HTTP/1.1 201
< Server: nginx/1.13.0
< Date: Thu, 25 May 2017 06:54:59 GMT
< Content-Length: 0
< Connection: keep-alive
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
< Location: http://localhost/posts/test-post-2/comments/3
<
* Connection #0 to host localhost left intact

Create another comment.

curl -v  http://localhost/posts/test-post-2/comments -X POST -H "X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8" -H "Content-Type:application/json" -d "{ \"content\":\"conmment content of post, another comment\"}"
Note: Unnecessary use of -X or --request, POST is already inferred.
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
*   Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to localhost (127.0.0.1) port 80 (#0)
> POST /posts/test-post-2/comments HTTP/1.1
> Host: localhost
> User-Agent: curl/7.54.0
> Accept: */*
> X-Auth-Token: 8b185a90-37db-444a-832b-6cbcd6db6df8
> Content-Type:application/json
> Content-Length: 56
>
* upload completely sent off: 56 out of 56 bytes
< HTTP/1.1 201
< Server: nginx/1.13.0
< Date: Thu, 25 May 2017 06:55:21 GMT
< Content-Length: 0
< Connection: keep-alive
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
< Location: http://localhost/posts/test-post-2/comments/4
<
* Connection #0 to host localhost left intact

Verify the comments.

curl -v  http://localhost/posts/test-post-2/comments  -H "Accpet:application/json"
* timeout on name lookup is not supported
*   Trying ::1...
* TCP_NODELAY set
* connect to ::1 port 80 failed: Connection refused
*   Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to localhost (127.0.0.1) port 80 (#0)
> GET /posts/test-post-2/comments HTTP/1.1
> Host: localhost
> User-Agent: curl/7.54.0
> Accept: */*
> Accpet:application/json
>
< HTTP/1.1 200
< Server: nginx/1.13.0
< Date: Thu, 25 May 2017 06:55:35 GMT
< Content-Type: application/json;charset=UTF-8
< Transfer-Encoding: chunked
< Connection: keep-alive
< X-Content-Type-Options: nosniff
< X-XSS-Protection: 1; mode=block
< Cache-Control: no-cache, no-store, max-age=0, must-revalidate
< Pragma: no-cache
< Expires: 0
< X-Frame-Options: DENY
<
{
  "content" : [ {
    "content" : "conmment content of post, another comment",
    "id" : 4,
    "createdDate" : "2017-05-25T06:55:22",
    "author" : {
      "username" : "user"
    }
  }, {
    "content" : "conmment content of post 2",
    "id" : 3,
    "createdDate" : "2017-05-25T06:54:59",
    "author" : {
      "username" : "user"
    }
  } ],
  "pageable" : {
    "sort" : {
      "sorted" : true,
      "unsorted" : false
    },
    "pageSize" : 10,
    "pageNumber" : 0,
    "offset" : 0,
    "paged" : true,
    "unpaged" : false
  },
  "last" : true,
  "totalElements" : 2,
  "totalPages" : 1,
  "sort" : {
    "sorted" : true,
    "unsorted" : false
  },
  "numberOfElements" : 2,
  "first" : true,
  "size" : 10,
  "number" : 0
}* Connection #0 to host localhost left intact

Testing Microservice

openssl req -x509 -nodes -days 365 -newkey rsa:2048 -keyout ssl/nginx.key -out ssl/nginx.crt

Create a private Docker Registry

• https://github.com/boot2docker/boot2docker/pull/1195
• docker-machine scp certfile default:ca.crt
• docker-machine ssh default
• sudo mv ~/ca.crt /etc/docker/certs.d/hostname/ca.crt
• exit
• docker-machine restart

openssl req -newkey rsa:4096 -nodes -sha256 -keyout certs/domain.key
-x509 -days 356 -out certs/domain.crt

Deployment on Docker Swarm

Use Docker Machine to create multi nodes.

In order to demonstrate running this project in Swarm mode, we will create two managers and three workers.

$ docker-machine create -d virtualbox --engine-registry-mirror https://docker.mirrors.ustc.edu.cn manager1
$ docker-machine create -d virtualbox --engine-registry-mirror https://docker.mirrors.ustc.edu.cn manager2
$ docker-machine create -d virtualbox --engine-registry-mirror https://docker.mirrors.ustc.edu.cn worker1
$ docker-machine create -d virtualbox --engine-registry-mirror https://docker.mirrors.ustc.edu.cn worker2
$ docker-machine create -d virtualbox --engine-registry-mirror https://docker.mirrors.ustc.edu.cn worker3

List all docker machines you just created.

$ docker-machine ls
NAME       ACTIVE   DRIVER       STATE     URL                         SWARM   DOCKER        ERRORS
manager1   -        virtualbox   Running   tcp://192.168.99.101:2376           v17.05.0-ce
manager2   -        virtualbox   Running   tcp://192.168.99.102:2376           v17.05.0-ce
worker1    -        virtualbox   Running   tcp://192.168.99.103:2376           v17.05.0-ce
worker2    -        virtualbox   Running   tcp://192.168.99.104:2376           v17.05.0-ce
worker3    -        virtualbox   Running   tcp://192.168.99.105:2376           v17.05.0-ce

Switch to machine manager1.

eval "$(docker-manager env manager1)"

Try to initialize a Docker Swarm host.

$ docker swarm init --listen-addr 192.168.99.101 --advertise-addr 192.168.99.101
Swarm initialized: current node (t36lxk020fasw5tdes4gm9ucf) is now a manager.

To add a worker to this swarm, run the following command:

    docker swarm join \
    --token SWMTKN-1-10bwwj2u6erepp9oc0qlkwao4o79vogifon51qkhdqfsl7zkkd-810eddvkzt2g8vvxb4gul4pnb \
    192.168.99.101:2377

To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.

We want to add manager2 as manager in this swarm. Follow the above info. Execute docker swarm join-token manager, it will show the guide to add more managers.

$ docker swarm join-token manager
To add a manager to this swarm, run the following command:

    docker swarm join \
    --token SWMTKN-1-10bwwj2u6erepp9oc0qlkwao4o79vogifon51qkhdqfsl7zkkd-4xus5y6wa7a4ass0f5bt20pym \
    192.168.99.101:2377

Let us switch to machine manager2.

eval "$(docker-machine env manager2)"

Copy and paste the docker swarm join command lines and execute it.

$ docker swarm join \
     --token SWMTKN-1-10bwwj2u6erepp9oc0qlkwao4o79vogifon51qkhdqfsl7zkkd-4xus5y6wa7a4ass0f5bt20pym \
     192.168.99.101:2377
This node joined a swarm as a manager.

Switch to worker1, worker2, and worker3, join this swarm as a worker.

    docker swarm join \
    --token SWMTKN-1-10bwwj2u6erepp9oc0qlkwao4o79vogifon51qkhdqfsl7zkkd-810eddvkzt2g8vvxb4gul4pnb \
    192.168.99.101:2377

Switch to any manager machine, and you can show all running nodes.

$ docker node ls
ID                            HOSTNAME            STATUS              AVAILABILITY        MANAGER STATUS
9d07by6czpem6hx55ke3ks1v1     manager2            Ready               Active              Reachable
er9klqvww0kdwyfaxr5f7n15l     worker1             Ready               Active
hsmaugexj4l7p5ighl9nega8q     worker2             Ready               Active
lknqw5dg5jyxw3j2camcpnb0v *   manager1            Ready               Active              Leader
ovqfs7ymrgbeyfqu8db8n6apc     worker3             Ready               Active

Switch to any manager machine, deploy all service via docker stack command.

docker stack deploy -c docker-stack.yml blogapp

The services will be scheduled to deploy in this swarm.

The docker-stack.yml file includes a visualizer service to visualize all services. It can be accessed via http://<any manager ip>:8080, you will see the deployment progress.

#curl http://192.168.99.102/user -u user:test123
{"roles":["ROLE_USER"],"name":"user"}

Remove this stack by the following command.

docker stack rm blogapp

Microservice deployment on Kubernetes