I’m looking for some guidance on the best way of writing integration (ie tests for entire Spring Boot Application) for Spring Cloud Task.
Based on existing documentation and samples I see two approaches for this:
1) Use the standard #SpringBootTest with #TestPropertySource(properties = {"spring.cloud.task.closecontext_enable=false"}
as described here
http://docs.spring.io/spring-cloud-task/docs/1.2.0.M2/reference/htmlsingle/#_writing_your_test
This seems to allow effectively only one test per test class as the task is run when the spring context is initialized ie once per test class. However
#Autowiring of beans in the context into the test class should work to eg to check the results of the task, or examine the state of the task repository.
2) Use SpringApplication.run(MyTaskApplication.class, myArguments); in each test method as in the example here
https://github.com/spring-cloud/spring-cloud-task/blob/master/spring-cloud-task-samples/batch-job/src/test/java/io/spring/BatchJobApplicationTests.java
This allows me to write multiple tests in the test class each with potentially different spring properties or batch job parameters.
The main problem I have with either approach that I can’t see how to get access eg #Autowire to beans in the context such as JdbcTemplate (eg to insert test input data for a job into an embedded db) or RestTemplate (to set up expectations using MockRestServiceServer)
after these beans are created but BEFORE the task is run - is this possible? If not it’s hard to see how to write meaningful integration tests for tasks.
What Ive done for now is a variation on approach (2) above (ie I can run the task more than once / have multiple tests in the same test class)
I'm using
SpringApplication application = new SpringApplication(new Object[] {MyTaskApplication.class, TestConfig.class});
TestConfig is defined with #TestConfiguration in the test class and contains mock beans etc overriding the actual beans
I then use
application.addListeners()
to add a ContextRefreshedEventListener which allows me to set expectations on mocks (or execute jdbc calls) after the beans are created but before the task is run.
(I have a generic Listener class that allows me to pass in the behaviour as a lambda or method reference per bean)
Then run the task with
application.run(args);
(can use different args in different tests)
I can also pass "--spring.cloud.task.closecontext_enable=false" as an argument to keep the application open if i want to verify mocks / db state after the test runs. In this case I close it manually at the end of the test.
If this seems a sensible approach it might be useful if Spring Cloud Task itself provided some sort of generic listener or hook to allow the setting of test state between bean creation and task execution.
Related
I have a cucumber test setup with spring boot. There are a large number of integration tests which take a while to run. Because they share a database, the new threaded mode in cucumber 4+ does not work (as expected).
Ideally this would work in Junit also.
For each test, I would like to create a dynamic datasource with a new database/datasource instance that the test can use independently of others, allowing it to run multithreaded (and use the 12 cores I have available).
I have tried #Scope("cucumber-glue") and #Scope("prototype") on the DataSource bean, but this results in org.springframework.beans.factory.BeanCurrentlyInCreationException: Error creating bean with name 'dataSource': Requested bean is currently in creation: Is there an unresolvable circular reference?.
If I use the prototype scope, the bean creation method gets called each time, but gives this error, as does the glue scope.
I have also added #DirtiesContext(classMode = DirtiesContext.ClassMode.BEFORE_EACH_TEST_METHOD) to classes and methods, without the scope, but this seems to do nothing.
Is there a way I can either:
1. create a new datasource for each test and have hibernate create the tables?
2. pool a set of datasources that could be used by tests?
3. Populate/reinitialize the context accordingly?
A side consequence of this is that I don't believe my context is getting recreated properly between tests for other instances of the scoping.
#Configuration
public class H2DynamicDataSource {
#Autowired
private Environment env;
#Bean
#Scope("prototype")
public DataSource dataSource() {
return new EmbeddedDatabaseBuilder()
.setType(EmbeddedDatabaseType.H2)
.build();
}
}
Cheers
R
Hopefully you've solved this.
I just went through something similar and felt like cruising around SlackOverflow to see if anyone was doing something similar (or perhaps asking about how to do something like this) and saw this post. Figured this would be a good place to drop this information in case anyone else tries to go down this road.
I think your asking two questions:
How to get multithreading working with Cucumber and Junit Cucumber?
multithreading works great but you are constrained to a single lane
of test execution by junit if your using either the Cucumber or
SpringJUnit4ClassRunner runner junit runner classes. (I suspect this is why #CucumberOptions doesnt contain a thread arg. It's useless with the current Cucumber runner implementation.)
In order to get this working I had to write my own test runner that
would defer a the entire test execution to Cucumbers Runtime
object. I used a custom cucumber plugin to cache the results and
then relay those results back to junit when asked to 'run' the
individual tests. With Cucumber behind the wheel I was able to allow
any arbitrary number of threads for my tests.
How to ensure a new spring datasource for each test?
If you're unfamiliar with how Spring cache's contexts for tests you should have a look through the documentation here https://docs.spring.io/spring/docs/current/spring-framework-reference/testing.html#testcontext-ctx-management-caching but the tl;dr: Spring's TestContextManager(this is whats under the hood of Springs junit runner) will create and store a spring context in a map keyed by your test configuration. DirtiesContext will force the context to reload.. but otherwise two tests with the same parent class are effectively guaranteed to execute in the same test context. If your datasource is a spring datasource that is initialized during boot.. you definitely need to refresh this context between tests.
Putting those two concepts together.. Cucumber-Spring provides a way to create a new context (consistent with cucumbers new 'world' per test design) for every test but it does so by disregarding any existing spring contexts or data contained therein. This may actually be helpful to you if you can trust Cucumber-Spring to correctly stand up your datasource for you.. but in my situation we had a bunch of issues using this fake context and really needed to pull objects from the default spring context. In my case I had to incorporate Springs TestContextManager into my custom plugin AND write my own Cucumber BackendSupplierimplementation to hijack Cucumbers dependency injection mechanism (effectively replacing Cucumber-Spring)
All in all, what your trying to do is a major PITA and I hope the Cucumber folks make this easier at some point.. but what your trying to do is definitely possible. Hope you got it working!
How we can programmatically configure Spring Boot to define new values to the spring.config.name and spring.config.location properties when running JUnit tests?
For example, if we would like to define these properties when running the application itself we could use something like (in Kotlin):
fun main(args: Array<String>) {
// SpringApplication.run(Application::class.java, *args)
val applicationContext = SpringApplicationBuilder(Application::class.java)
.properties(
"""spring.config.name:
${getSpringConfigNames()}
""",
"""spring.config.location:
${getSpringConfigLocationPaths()}
"""
)
.build()
.run(*args)
// val environment = applicationContext.getEnvironment()
}
But I wasn't able to find a way to configure this to use in the JUnit tests.
Edit
There is a complication here because of an spring boot limitation.
I would like to use an entire folder and its subfolders as valid locations to search for configuration files (so, for example, we could have folders for specific environments, databases, third-parties, and so on).
When running the application this was possible creating a method, in this case getSpringConfigLocationPaths(). And this method create a comma separated list with all folder inside the "main" folder.
For example, for the main folder src/main/resources/configuration it will output:
src/main/resources/configuration,
src/main/resources/configuration/environments,
src/main/resources/configuration/environments/development,
src/main/resources/configuration/environments/staging,
src/main/resources/configuration/environments/testing,
src/main/resources/configuration/environments/production,
src/main/resources/configuration/environments/common
How could we solve this situation when using JUnit tests and Spring Boot?
Unfortunately Spring Boot doesn't allow something like src/main/resources/configuration/**/*.
Because we have organized the system with several properties files on different subfolders we need to find a way to dinamically consider them.
I am using latest Spring Boot 2.2.0 and from my experience both #TestPropertySource and #SpringBootTest annotations can do the job because they have properties attribute.
So, you can do something like this:
#TestPropertySource(properties = ["spring.config.location=classpath:dev/", "spring.config.name=custom-app-name"]
#TestConfiguration
class DevTestCfg {} // this will make tests to look for configs in resources/dev/custom-app-name.properties
Also notice that there is a spring.config.additional-location property if you want your properties to be loaded from multiple locations.
The only problem here is that values in properties attribute must be constant.
But you can create multiple configurations for each environment and put corresponding #Profile("envName") on each configuration class. Then run your tests with different -Dspring.profiles.active and corresponding test configuration should be automatically picked up.
The tests that run spring boot should be carefully designed,
There is a whole testing framework for spring boot tests, so obviously consider using this framework.
When it comes to configuration management, I suggest considering the following:
There are two types of tests basically:
Tests that load a concrete specific configuration (set of beans), for example if you want to test only a DAO, you load a configuration for this dao.
In this case, the configuration is something that should be "tailored" to the needs of a specific test, and no "full" configuration is required.
For example, if the microservice contains a configuration for a database (user, password, schema, etc) and for, say, messaging management, there is no need to specify a configuration of a messaging system when testing a DAO, messaging beans won't be loaded anyway.
Usually, the test of this "type" will look like this:
#SpringBootTest(classes = {RelationalDbDaoConfiguration.class})
public class MyDaoTest {
}
If you don't have a configuration for your needs you can use #MockBean to mock unnecessary beans or even create a custom configuration in src/test/java so that it will be only in test classpath. It makes sense to use #TestConfiguration but it's beyond the scope of the question.
Now in order to load the configuration for db only, the are many options, to name a few:
#ActiveProfiles("dao") on a test class + putting "application-dao.properties/yaml" into the src/test/resources or src/test/resources/config
Use #TestPropertySource(locations = "classpath:whatever.properties") on test
Create a special "DbProperties" bean and initialize it programmatically in spring, it can make sense when you know some details about the context in which the test runs only during the actual test execution (for example, if you start a database before the test and the port is created dynamically, but its really a fairly advanced setup and is beyond the scope of this question) + the data source bean can read these properties
Use #SpringBootTest's properties attribute to provide 'fine-grained' properties definitions
Kind of obvious, but I'll mention it anyway: put application.properties in src/test/resources it will override regular configurations
The second type of tests is when you load the "entire" microservice, usually, these are tests that do not have "classes" parameter in #SpringBootTest annotation
#SpringBootTest // note, no actual configurations specified
public class MyMicroserviceTest {
...
}
Now, this definitely requires to specify a whole set of configurations, although the techniques for actually specifying these configurations are still applicable (just the content of configuration files will be different).
I do not suggest the usage of spring.config.location during the test, because this means that the test depends on some external resource, which makes the whole setup even more complicated.
If it's XML driven configuration,
#ContextConfiguration(locations = "/app-context.xml")
If it's annotation driven by configuration classes,
#ContextConfiguration(classes = {AppCOnfig::class, AnotherCOnfig::class}
These would be defined on the class level on the unit test class you run.
Further, if you have profiles for Junit to consider,
#ActiveProfiles("myProfile") would be added to the test class.
What's the recommended way to run a spring boot test where only the one subject under test is configured in the context.
If I annotate the test with
#RunWith(SpringRunner.class)
#SpringBootTest(properties = "spring.profiles.active=test")
#ContextConfiguration(classes = MyTestBean.class)
Then it seems to work - the test passes, the context starts quickly and seems to only contain the bean that I want. However, this seems like an incorrect use of the #ContextConfiguration(classes = MyTestBean.class) annotation. If I understand correctly the class that I reference is supposed to be a Configuration class, not a regular spring service bean or component for example.
Is that right? Or is this indeed a valid way to achieve this goal? I know there are more complex examples like org.springframework.boot.test.autoconfigure.json.JsonTest which use #TypeExcludeFilters(JsonExcludeFilter.class) to control the context - but this seems overkill for my use case. I just want a context with my one bean.
Clarification
I know that I can just construct the one bean I am testing as a POJO without a spring context test and remove the three annotations above. But in my precise use case I am actually reliant on some of the configuration applied to the context by settings in the application-test.properties file - which is why I've made this a Spring Boot test with a profile set. From my perspective this isn't a plain unit test of a single class in isolation of the spring context configuration - the test is reliant on certain configuration being applied (which is currently provided by the spring boot app properties). I can indeed just test the components as a POJO by creating a new instance outside of a spring context, I'm using constructor injection making the providing of necessary dependencies simple but the test does rely on things like the log level (the test actually makes assertions on certain logs being produced) which requires that the log level is set correctly (which is currently being done via logging.level.com.example=DEBUG in a properties file which sets up the spring context).
For starters, reading the documentation first (e.g., the JavaDoc linked below in this answer) is a recommend best practice since it already answers your question.
If I understand correctly the class that I reference is supposed to be
a Configuration class, not a regular spring service bean or
component for example.
Is that right?
No, that's not completely correct.
Classes provided to #ContextConfiguration are typically #Configuration classes, but that is not required.
Here is an excerpt from the JavaDoc for #ContextConfiguration:
Annotated Classes
The term annotated class can refer to any of the following.
A class annotated with #Configuration
A component (i.e., a class annotated with #Component, #Service, #Repository, etc.)
A JSR-330 compliant class that is annotated with javax.inject annotations
Any other class that contains #Bean-methods
Thus you can pass any "annotated class" to #ContextConfiguration.
Or is this indeed a valid way to achieve this goal?
It is in fact a valid way to achieve that goal; however, it is also a bit unusual to load an ApplicationContext that contains a single user bean.
Regards,
Sam (author of the Spring TestContext Framework)
It is definitely a reasonable and normal thing to only test a single class in a unit test.
There is no problem including just one single bean in your test context. Really, a #Configuration is (typically) just a collection of beans. You could hypothetically create a #Configuration class just with MyTestBean, but that would really be unnecessary, as you can accomplish doing the same thing listing your contextual beans with #ContextConfiguration#classes.
However, I do want to point out that for only testing a single bean in a true unit test, best practice ideally leans towards setting up the bean via the constructor and testing the class that way. This is a key reason why the Spring guys recommend using constructor vs. property injection. See the section entitled Constructor-based or setter-based DI of this article, Oliver Gierke's comment (i.e. head of Spring Data project), and google for more information. This is probably the reason you're getting a weird feeling about setting up the context for the one bean!
You can also use ApplicationContextRunner to create your context using a test configuration of your choice (even with one bean if you like, but as other people have already mentioned for one bean it's more reasonable to use the constructor the classical way without using any spring magic).
What I like this way of testing is the fact that test run very fast since you don't load all the context. This method is best used when the tested bean doesn't have any Autowired dependencies otherwise it's more convenient to use #SpringBootTest.
Below is an example that illustrates the way you can use it to achieve your goal:
class MyTest {
#Test
void test_configuration_should_contains_my_bean() {
new ApplicationContextRunner()
.withUserConfiguration(TestConfiguration.class)
.run(context -> {
assertThat(context.getBean(MyTestBean.class)).isNotNull();
});
}
#Configuraiton
public static class TestConfiguration {
#Bean
public MyTestBean myTestBean(){
new MyTestBean();
}
}
}
I have created a couple unittest classes in the same package. All these classes have exactly one testcase and have the same annotation as shown below:
#RunWith(SpringJUnit4ClassRunner.class)
#ContextConfiguration(classes= {TestConfig.class} )
When I run these testcases, I want each testcase be run in its own application context. But it seems all the testcases in the same package share one single application context, be it run from maven command line, or in Eclipse select the package to run as junit.
If I duplicate the TestConfig wiht names like TestConfig1, TestConfig2, etc. and annotate different test class with different TestConfig class, then each test will run in its own context instance.
Is there other elegant method to achieve this?
Thanks a lot.
This is the default behavior of Spring test. Spring test default caches the application context across the tests. This decreases the test execution time.
I dont know about your use case. If u have any where u are dirtying the application context(changing the state of the beans that are managed by springs that is affecting the subsequent tests ) then u can go for annotating the test method with #DirtiesContext. Spring test reloads the context for these methods. See below link how to use DirtiesContext.
https://docs.spring.io/spring/docs/current/spring-framework-reference/html/integration-testing.html#testcontext-ctx-management-caching
Use this feature carefully whenever needed only as this may shoot up your test execution time exponentialy.
I have a Spring Boot project for which one of the developers has written Integration tests against an in-memory DB. Here is some background before I put my question :
We have following Maven modules :
1) web-persistence : which store all entities and interfaces and corresponding implementations. All these classes are in package com.mycompany.persistence.
2) web-services : which stores all our rest controllers and spring security related classes. These are all in package com.mycompany.services
3) web-services module has maven module dependency on web-persistence module.
4) In web-services module we have defined a service UserContextServiceImpl which is implementation of interface IUserContextService. This service basically encapsulates Spring's SecurityContextHolder functionality by exposing only those parts of a current user related information which is needed through its methods so that everywhere we don't have to use SecurityContextHolder. The implementation class is annotated with #Service("userContextService").
5) Interface IUserContextService is in persistence module. There is another class AppDependecyHeper in persistence module which implements ApplicationContextAware and returns UserContextService's concrete instance by calling appContext.getBean('userContextService') where appContext is holding ApplicationContext when application is initialized.
6) Now we don't want to expose Spring Security classes in persistence module(for reasons which are out of context here) and hence above arrangement of getting current user information through a service by calling getBean() method of applicationContext in persistence layer class AppDependencyHelper. This information is then used to update audit fields createdBy and modifiedBy using an EntityListener.
7) Actual application uses only single ApplicationContext where both persistence layer classes and web controllers are loaded in one single spring context.
8) Everything works fine when application runs. However when our integration test Runs the call appContext.getBean('userContextService') fails and throws NosuchBeanDefinitionException as it is unable to find the bean with name 'userContextService'
9) Now finally the code of our integration Test (only giving relevant details) which is located in com.mycompany.persistence.embeddeddb package:
#Runwith(SpringJUnit4ClassRunner.class)
#ConfigurationContext
class MyEntityTest{
#Configuration
#ComponentScan("com.mycompany.persistence")
public static class Config{
.....
}
.....//code which eventually gives call to
AppDependencyHelper's method which in turn tries to retrieve userContextService Bean.
}
Now the questions :
1) Why it fails to retrieve the bean? I have tried addingcom.mycompany.services like below
#ComponentScan({"com.mycompany.persistence","com.mycompany.services"})
but to no avail.
2) What I am able to understand by reading whatever I got is that #ContextConfiguration needs to be provided with either XML files or Annotated classes or WebApplicationInitializers(in my case as we are having only single ApplicationContext) to create an ApplicationContext from #Configuration classes. However I cannot do that as maven starts complaining about circular dependency which is rightly so as my #Service annotated class UserContextServiceImpl is in web-services module which is dependent on web-persistence module already where the test case class is written.
3) The only solution I can think of is moving all integrations test classes to web-services module so that I can tell #ConfigurationContext all the classes from which it should create an ApplicationContext.Am I right?
Last point : The ApplicationContext class is of type GenericApplciationContext when I run tests. When I run application its obviously AnnotatedEmbeddedWebApplicationContext
please let me know if there is any solution to this problem?
From what I'm guessing you are trying to use a component from your web-service module in your integration tests. If that's the case, than yes, you should move your integration test cases to your web-service module - as you've already suggested in your question.
EDIT: You should really rethink your design, because you have a circular dependency problem here: Your web-service module depends on web-persistence, but the implementation of your IUserContextService in web-persistence is found in web-service.
You should probably move your IUserContextService (and all your integration tests) to your web-service package.