The English version of quarkus.io is the official project site. Translated sites are community supported on a best-effort basis.

Tips for writing native applications

This guide contains various tips and tricks for getting around problems that might arise when attempting to run Java applications as native executables.

Note that we differentiate two contexts where the solution applied might be different:

  • in the context of an application, you will rely on configuring the native-image configuration by tweaking your pom.xml;

  • in the context of an extension, Quarkus offers a lot of infrastructure to simplify all of this.

Please refer to the appropriate section depending on your context.

Supporting native in your application

GraalVM imposes a number of constraints and making your application a native executable might require a few tweaks.

Including resources

By default, when building a native executable, GraalVM will not include any of the resources that are on the classpath into the native executable it creates. Resources that are meant to be part of the native executable need to be configured explicitly.

Quarkus automatically includes the resources present in META-INF/resources (the web resources) but, outside this directory, you are on your own.

Note that you need to be extremely careful here as anything in META-INF/resources will be exposed as static web resources. So this directory is not a shortcut for "let’s automatically include these resources in the native executable" and should only be used for static web resources.

Other resources should be declared explicitly.

Using the quarkus.native.resources.includes configuration property

To include more resources in the native executable, the easiest way is to use the quarkus.native.resources.includes configuration property, and its counterpart to exclude resources quarkus.native.resources.excludes.

Both configuration properties support glob patterns.

For instance, having the following properties in your application.properties:

quarkus.native.resources.includes=foo/**,bar/**/*.txt
quarkus.native.resources.excludes=foo/private/**

will include:

  • all files in the foo/ directory and its subdirectories except for files in foo/private/ and its subdirectories,

  • all text files in the bar/ directory and its subdirectories.

Using a configuration file

If globs are not sufficiently precise for your use case and you need to rely on regular expressions, or if you prefer relying on the GraalVM infrastructure, you can also create a resource-config.json JSON file defining which resources should be included. This, and other native image configuration files, should be placed under the src/main/resources/META-INF/native-image/<group-id>/<artifact-id> folder. This way they will be automatically parsed by the native build, without additional configuration.

Relying on the GraalVM infrastructure means that you are responsible for keeping the configuration file up to date as new Mandrel and GraalVM versions are released.

Please also note that the resource-config.json file will be overwritten by Quarkus if placed directly under src/main/resources/META-INF/native-image/ as Quarkus generates its own configuration file in this directory.

An example of such a file is the following:

{
  "resources": [
    {
      "pattern": ".*\\.xml$"
    },
    {
      "pattern": ".*\\.json$"
    }
  ]
}

The patterns are valid Java regexps. Here we include all the XML files and JSON files into the native executable.

For more information about this topic, see the GraalVM Accessing Resources in Native Image guide.

Registering for reflection

When building a native executable, GraalVM operates with a closed world assumption. It analyzes the call tree and removes all the classes/methods/fields that are not used directly.

The elements used via reflection are not part of the call tree so they are dead code eliminated (if not called directly in other cases). To include these elements in your native executable, you need to register them for reflection explicitly.

This is a very common case as JSON libraries typically use reflection to serialize the objects to JSON:

    public class Person {
        private String first;
        private String last;

        public String getFirst() {
            return first;
        }

        public void setFirst(String first) {
            this.first = first;
        }

        public String getLast() {
            return last;
        }

        public void setValue(String last) {
            this.last = last;
        }
    }

    @Path("/person")
    @Produces(MediaType.APPLICATION_JSON)
    @Consumes(MediaType.APPLICATION_JSON)
    public class PersonResource {

        private final Jsonb jsonb;

        public PersonResource() {
            jsonb = JsonbBuilder.create(new JsonbConfig());
        }

        @GET
        public Response list() {
            return Response.ok(jsonb.fromJson("{\"first\":  \"foo\", \"last\":  \"bar\"}", Person.class)).build();
        }
    }

If we were to use the code above, we would get an exception like the following when using the native executable:

Exception handling request to /person: org.jboss.resteasy.spi.UnhandledException: jakarta.json.bind.JsonbException: Can't create instance of a class: class org.acme.jsonb.Person, No default constructor found

or if you are using Jackson:

com.fasterxml.jackson.databind.exc.InvalidDefinitionException: No serializer found for class org.acme.jsonb.Person and no properties discovered to create BeanSerializer (to avoid exception, disable SerializationFeature.FAIL_ON_EMPTY_BEANS)

An even nastier possible outcome could be for no exception to be thrown, but instead the JSON result would be completely empty.

There are two different ways to fix this type of issues.

Using the @RegisterForReflection annotation

The easiest way to register a class for reflection is to use the @RegisterForReflection annotation:

@RegisterForReflection
public class MyClass {
}

If your class is in a third-party jar, you can do it by using an empty class that will host the @RegisterForReflection for it.

@RegisterForReflection(targets={ MyClassRequiringReflection.class, MySecondClassRequiringReflection.class})
public class MyReflectionConfiguration {
}

Note that MyClassRequiringReflection and MySecondClassRequiringReflection will be registered for reflection but not MyReflectionConfiguration.

This feature is handy when using third-party libraries using object mapping features (such as Jackson or GSON):

@RegisterForReflection(targets = {User.class, UserImpl.class})
public class MyReflectionConfiguration {

}

Note: By default the @RegisterForReflection annotation will also registered any potential nested classes for reflection. If you want to avoid this behavior, you can set the ignoreNested attribute to true.

Using a configuration file

You can also use a configuration file to register classes for reflection, if you prefer relying on the GraalVM infrastructure.

Relying on the GraalVM infrastructure means that you are responsible for keeping the configuration file up to date as new Mandrel and GraalVM versions are released.

As an example, in order to register all methods of class com.acme.MyClass for reflection, we create reflect-config.json (the most common location is within src/main/resources)

[
  {
    "name" : "com.acme.MyClass",
    "allDeclaredConstructors" : true,
    "allPublicConstructors" : true,
    "allDeclaredMethods" : true,
    "allPublicMethods" : true,
    "allDeclaredFields" : true,
    "allPublicFields" : true
  }
]

For more information about the format of this file, see the GraalVM Reflection in Native Image guide.

The final order of business is to make the configuration file known to the native-image executable. To do that, place the configuration file under the src/main/resources/META-INF/native-image/<group-id>/<artifact-id> folder. This way they will be automatically parsed by the native build, without additional configuration.

Registering for proxy

Analogous to @RegisterForReflection, you can use @RegisterForProxy to register interfaces for dynamic proxy:

@RegisterForProxy
public interface MyInterface extends MySecondInterface {
}

Note that MyInterface and all its super interfaces will be registered.

Also, in case the interface is in a third-party jar, you can do it by using an empty class that will host the @RegisterForProxy for it.

@RegisterForProxy(targets={MyInterface.class, MySecondInterface.class})
public class MyReflectionConfiguration {
}

Note that the order of the specified proxy interfaces is significant. For more information, see Proxy javadoc.

Delaying class initialization

By default, Quarkus initializes all classes at build time.

There are cases where the initialization of certain classes is done in a static block needs to be postponed to runtime. Typically, omitting such configuration would result in a runtime exception like the following:

Error: No instances are allowed in the image heap for a class that is initialized or reinitialized at image runtime: sun.security.provider.NativePRNG
Trace: object java.security.SecureRandom
method com.amazonaws.services.s3.model.CryptoConfiguration.<init>(CryptoMode)
Call path from entry point to com.amazonaws.services.s3.model.CryptoConfiguration.<init>(CryptoMode):

Another common source of errors is when the image heap taken by GraalVM contains a Random/SplittableRandom instance:

Error: com.oracle.graal.pointsto.constraints.UnsupportedFeatureException: Detected an instance of Random/SplittableRandom class in the image heap. Instances created during image generation have cached seed values and don't behave as expected.

Which is more often than not caused by Quarkus initializing at build time a class with a static Random/SplittableRandom field, causing this particular instance to be tentatively included in the image heap.

You can find detailed information about this Random/SplittableRandom issue in this blog post.

In these cases, delaying the infringing class initialization at runtime might be the solution and, to achieve that, you can use the --initialize-at-run-time=<package or class> configuration knob.

It should be added to the native-image configuration using the quarkus.native.additional-build-args configuration property as shown in the examples above.

For more information, see the GraalVM Class Initialization in Native Image guide.

When multiple classes or packages need to be specified via the quarkus.native.additional-build-args configuration property, the , symbol needs to be escaped. An example of this is the following:

quarkus.native.additional-build-args=--initialize-at-run-time=com.example.SomeClass\\,org.acme.SomeOtherClass

and in the case of using the Maven configuration instead of application.properties:

<quarkus.native.additional-build-args>--initialize-at-run-time=com.example.SomeClass\,org.acme.SomeOtherClass</quarkus.native.additional-build-args>

Managing Proxy Classes

While writing native application you’ll need to define proxy classes at image build time by specifying the list of interfaces that they implement.

In such a situation, the error you might encounter is:

com.oracle.svm.core.jdk.UnsupportedFeatureError: Proxy class defined by interfaces [interface org.apache.http.conn.HttpClientConnectionManager, interface org.apache.http.pool.ConnPoolControl, interface com.amazonaws.http.conn.Wrapped] not found. Generating proxy classes at runtime is not supported. Proxy classes need to be defined at image build time by specifying the list of interfaces that they implement. To define proxy classes use -H:DynamicProxyConfigurationFiles=<comma-separated-config-files> and -H:DynamicProxyConfigurationResources=<comma-separated-config-resources> options.

Solving this issue requires creating a proxy-config.json file under the src/main/resources/META-INF/native-image/<group-id>/<artifact-id> folder. This way the configuration will be automatically parsed by the native build, without additional configuration. For more information about the format of this file, see the Dynamic Proxy Metadata in JSON documentation.

Modularity Benefits

During native executable build time GraalVM analyses the application’s call tree and generates a code-set that includes all the code it needs. Having a modular codebase is key to avoiding problems with unused or optional parts of your application, while at the same time reducing both native executable build times and size. In this section you will learn about the details behind the benefits of modularity for native applications.

When code is not modular enough, generated native executables can end up with more code than what the user needs. If a feature is not used and the code gets compiled into the native executable, this is a waste of native compilation time and memory usage, as well as native executable disk space and starting heap size. Even more problems arise when third party libraries or specialized API subsystems are used which cause native compilation or runtime errors, and their use is not modularised enough. A recent problem can be found in the JAXB library, which is capable of deserializing XML files containing images using Java’s AWT APIs. The vast majority of Quarkus XML users don’t need to deserialize images, so there shouldn’t be a need for users applications to include Java AWT code, unless they specifically configure Quarkus to add the JAXB AWT code to the native executable. However, because JAXB code that uses AWT is in the same jar as the rest of the XML parsing code, achieving this separation was rather complex and required the use of Java bytecode substitutions to get around it. These substitutions are hard to maintain and can easily break, hence they should be one’s last resort.

A modular codebase is the best way to avoid these kind of issues. Taking the JAXB/AWT problem above, if the JAXB code that dealt with images was in a separate module or jar (e.g. jaxb-images), then Quarkus could choose not to include that module unless the user specifically requested the need to serialize/deserialize XML files containing images at build time.

Another benefit of modular applications is that they can reduce the code-set that will need to get into the native executable. The smaller the code-set, the faster the native executable builds will be and the smaller the native executable produced.

The key takeaway point here is the following: Keeping optional features, particularly those that depend on third party libraries or API subsystems with a big footprint, in separate modules is the best solution.

How do I know if my application suffers from similar problems? Aside from a deep study of the application, finding usages of Maven optional dependencies is a clear indicator that your application might suffer from similar problems. These type of dependencies should be avoided, and instead code that interacts with optional dependencies should be moved into separate modules.

Enforcing Singletons

As already explained in the delay class initialization section, Quarkus marks all code to be initialized at build time by default. This means that, unless marked otherwise, static variables will be assigned at build time, and static blocks will be executed at build time too.

This can cause values in Java programs that would normally vary from one run to another, to always return a constant value. E.g. a static field that is assigned the value of System.currentTimeMillis() will always return the same value when executed as a Quarkus native executable.

Singletons that rely on static variable initialization will suffer similar problems. For example, imagine you have a singleton based around static initialization along with a REST endpoint to query it:

@Path("/singletons")
public class Singletons {

    @GET
    @Path("/static")
    public long withStatic() {
        return StaticSingleton.startTime();
    }
}

class StaticSingleton {
    static final long START_TIME = System.currentTimeMillis();

    static long startTime() {
        return START_TIME;
    }
}

When the singletons/static endpoint is queried, it will always return the same value, even after the application is restarted:

$ curl http://localhost:8080/singletons/static
1656509254532%

$ curl http://localhost:8080/singletons/static
1656509254532%

### Restart the native application ###

$ curl http://localhost:8080/singletons/static
1656509254532%

Singletons that rely on enum classes are also affected by the same issue:

@Path("/singletons")
public class Singletons {

    @GET
    @Path("/enum")
    public long withEnum() {
        return EnumSingleton.INSTANCE.startTime();
    }
}

enum EnumSingleton {
    INSTANCE(System.currentTimeMillis());

    private final long startTime;

    private EnumSingleton(long startTime) {
        this.startTime = startTime;
    }

    long startTime() {
        return startTime;
    }
}

When the singletons/enum endpoint is queried, it will always return the same value, even after the application is restarted:

$ curl http://localhost:8080/singletons/enum
1656509254601%

$ curl http://localhost:8080/singletons/enum
1656509254601%

### Restart the native application ###

$ curl http://localhost:8080/singletons/enum
1656509254601%

One way to fix it is to build singletons using CDI’s @Singleton annotation:

@Path("/singletons")
public class Singletons {

    @Inject
    CdiSingleton cdiSingleton;

    @GET
    @Path("/cdi")
    public long withCdi() {
        return cdiSingleton.startTime();
    }
}

@Singleton
class CdiSingleton {
    // Note that the field is not static
    final long startTime = System.currentTimeMillis();

    long startTime() {
        return startTime;
    }
}

After each restart, querying singletons/cdi will return a different value, just like it would in JVM mode:

$ curl http://localhost:8080/singletons/cdi
1656510218554%

$ curl http://localhost:8080/singletons/cdi
1656510218554%

### Restart the native application ###

$ curl http://localhost:8080/singletons/cdi
1656510714689%

An alternative way to enforce a singleton while relying static fields, or enums, is to delay its class initialization until run time. The nice advantage of CDI-based singletons is that your class initialization is not constrained, so you can freely decide whether it should be build-time or run-time initialized, depending on your use case.

Beware of common Java API overrides

Certain commonly used Java methods are overriden by user classes, e.g. toString, equals, hashCode…​etc. The majority of overrides do not cause problems, but if they use third party libraries (e.g. for additional formatting), or use dynamic language features (e.g. reflection or proxies), they can cause native image build to fail. Some of those failures might be solvable via configuration, but others can be more tricky to handle.

From a GraalVM points-to analysis perspective, what happens in these method overrides matters, even if the application does not explicitly call them. This is because these methods are used throughout the JDK, and all it takes is for one of those calls to be done on an unconstrained type, e.g. java.lang.Object, for the analysis to have to pull all implementations of that particular method.

Supporting native in a Quarkus extension

Supporting native in a Quarkus extension is even easier as Quarkus provides a lot of tools to simplify all this.

Everything described here will only work in the context of Quarkus extensions, it won’t work in an application.

Register reflection

Quarkus makes registration of reflection in an extension a breeze by using ReflectiveClassBuildItem, thus eliminating the need for a JSON configuration file.

To register a class for reflection, one would need to create a Quarkus processor class and add a build step that registers reflection:

public class SaxParserProcessor {

    @BuildStep
    ReflectiveClassBuildItem reflection() {
        // since we only need reflection to the constructor of the class, we can specify `false` for both the methods and the fields arguments.
        return new ReflectiveClassBuildItem(false, false, "com.sun.org.apache.xerces.internal.parsers.SAXParser");
    }

}

For more information about reflection in GraalVM, see the GraalVM Reflection in Native Image guide.

Including resources

In the context of an extension, Quarkus eliminates the need for a JSON configuration file by allowing extension authors to specify a NativeImageResourceBuildItem:

public class ResourcesProcessor {

    @BuildStep
    NativeImageResourceBuildItem nativeImageResourceBuildItem() {
        return new NativeImageResourceBuildItem("META-INF/extra.properties");
    }

}

For more information about GraalVM resource handling in native executables, see the GraalVM Accessing Resources in Native Image guide.

Delay class initialization

Quarkus simplifies things by allowing extensions authors to simply register a RuntimeInitializedClassBuildItem. A simple example of doing so could be:

public class S3Processor {

    @BuildStep
    RuntimeInitializedClassBuildItem cryptoConfiguration() {
        return new RuntimeInitializedClassBuildItem(CryptoConfiguration.class.getCanonicalName());
    }

}

Using such a construct means that a --initialize-at-run-time option will automatically be added to the native-image command line.

For more information about the --initialize-at-run-time option, see the GraalVM Class Initialization in Native Image guide.

Managing Proxy Classes

Very similarly, Quarkus allows extensions authors to register a NativeImageProxyDefinitionBuildItem. An example of doing so is:

public class S3Processor {

    @BuildStep
    NativeImageProxyDefinitionBuildItem httpProxies() {
        return new NativeImageProxyDefinitionBuildItem("org.apache.http.conn.HttpClientConnectionManager",
                "org.apache.http.pool.ConnPoolControl", "com.amazonaws.http.conn.Wrapped");
    }

}

Using such a construct means that a -H:DynamicProxyConfigurationResources option will automatically be added to the native-image command line.

For more information about Proxy Classes, see the GraalVM Configure Dynamic Proxies Manually guide.

Logging with Native Image

If you are using dependencies that require logging components such as Apache Commons Logging or Log4j and are experiencing a ClassNotFoundException when building the native executable, you can resolve this by excluding the logging library and adding the corresponding JBoss Logging adapter.

For more details please refer to the Logging guide.

Conteúdo Relacionado