Native image build failures on latest GraalVM with Random/SplittableRandom initialized at build time #14904

zakkak · 2021-02-08T14:30:27Z

Describe the bug
Quarkus TS fail with latest Graal (see https://github.com/oracle/graal/actions/runs/546718036)

In the upcoming release of GraalVM (21.1) instances of Random/SplittableRandom class won't be allowed in the native image heap. The corresponding patch can be seen here.

As a result, moving forward Quarkus needs to ensure that its core and extensions reinitialize any classes holding instances of Random/SplittableRandom class.

In order to find out which classes to reinitialize one needs to run:

./mvnw verify \
  -Dnative \
  -Dquarkus.native.additional-build-args="--trace-object-instantiation=java.security.SecureRandom" \
  -pl integration-tests/main

This will print out something like this:

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.  Object has been initialized by the org.apache.http.impl.auth.NTLMEngineImpl class initializer with a trace: 
        at java.security.SecureRandom.<init>(SecureRandom.java:323)
        at java.security.SecureRandom.getInstance(SecureRandom.java:383)
        at org.apache.http.impl.auth.NTLMEngineImpl.<clinit>(NTLMEngineImpl.java:103)
. Try avoiding to initialize the class that caused initialization of the object. The object was probably created by a class initializer and is reachable from a static field. You can request class initialization at image runtime by using the option --initialize-at-run-time=<class-name>. Or you can write your own initialization methods and call them explicitly from your main entry point.
Detailed message:
Trace: 
        at parsing org.apache.http.impl.auth.NTLMEngineImpl.access$000(NTLMEngineImpl.java:51)
Call path from entry point to org.apache.http.impl.auth.NTLMEngineImpl.access$000(): 
        at org.apache.http.impl.auth.NTLMEngineImpl.access$000(NTLMEngineImpl.java:51)
        at org.apache.http.impl.auth.NTLMEngineImpl$Type3Message.<init>(NTLMEngineImpl.java:1505)
        at org.apache.http.impl.auth.NTLMEngineImpl$Type3Message.<init>(NTLMEngineImpl.java:1474)
        at org.apache.http.impl.auth.NTLMEngineImpl.getType3Message(NTLMEngineImpl.java:181)
        at org.apache.http.impl.auth.NTLMEngineImpl.generateType3Msg(NTLMEngineImpl.java:2097)
        at org.apache.http.impl.auth.NTLMScheme.authenticate(NTLMScheme.java:142)
        at org.apache.http.impl.auth.AuthSchemeBase.authenticate(AuthSchemeBase.java:136)
        at org.apache.http.impl.auth.HttpAuthenticator.doAuth(HttpAuthenticator.java:233)
        at org.apache.http.impl.auth.HttpAuthenticator.generateAuthResponse(HttpAuthenticator.java:213)
        at org.apache.http.impl.execchain.MainClientExec.execute(MainClientExec.java:262)
        at org.apache.http.impl.execchain.ProtocolExec.execute(ProtocolExec.java:186)
        at org.apache.http.impl.client.InternalHttpClient.doExecute(InternalHttpClient.java:185)
        at org.apache.http.impl.client.CloseableHttpClient.execute(CloseableHttpClient.java:83)
        at org.apache.http.impl.client.CloseableHttpClient.execute(CloseableHttpClient.java:56)
        at org.jboss.resteasy.client.jaxrs.engines.ManualClosingApacheHttpClient43Engine.invoke(ManualClosingApacheHttpClient43Engine.java:268)
        at org.jboss.resteasy.client.jaxrs.internal.ClientInvocation.invoke(ClientInvocation.java:488)
        at org.jboss.resteasy.client.jaxrs.internal.ClientInvocation.lambda$executorSubmit$11(ClientInvocation.java:770)
        at org.jboss.resteasy.client.jaxrs.internal.ClientInvocation$$Lambda$2014/0x00000007c4cd7840.get(Unknown Source)
        at com.oracle.svm.core.jdk.SystemPropertiesSupport.initializeLazyValue(SystemPropertiesSupport.java:211)
        at com.oracle.svm.core.jdk.SystemPropertiesSupport.getProperty(SystemPropertiesSupport.java:164)
        at com.oracle.svm.core.jdk.Target_java_lang_System.getProperty(JavaLangSubstitutions.java:293)
        at com.oracle.svm.jni.JNIJavaCallWrappers.jniInvoke_VA_LIST:Ljava_lang_System_2_0002egetProperty_00028Ljava_lang_String_2_00029Ljava_lang_String_2(generated:0)

Which indicates that Quarkus should reinitialize NTLMEngineImpl at runtime.

Expected behavior
Tests should pass (at least most of them)

Actual behavior
Multiple integration tests fail.

Main
Hibernate Tenancy
Flyway
Redis Client
Hibernate Search ORM + Elasticsearch
Elasticsearch REST client
Elasticsearch REST high level client
Bouncy Castle
Bouncy Castle FIPS
Bouncy Castle JSSE
Elytron Security LDAP
more...
fail with:

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.  To see how this object got instantiated use --trace-object-instantiation=java.security.SecureRandom. The object was probably created by a class initializer and is reachable from a static field. You can request class initialization at image runtime by using the option --initialize-at-run-time=<class-name>. Or you can write your own initialization methods and call them explicitly from your main entry point.

To Reproduce

Build GraalVM master with:
mx --primary-suite-path substratevm --components='Native Image' build
export GRAALVM_HOME=/path/to/graal/sdk/latest_graalvm_home
./mvnw clean verify -Dnative -Dnative.surefire.skip -pl integration-tests/main

Environment (please complete the following information):

Output of uname -a or ver: Linux 5.10.10-200.fc33.x86_64
Output of java -version: 11.0.9+10-jvmci-21.0-b03
GraalVM version (if different from Java): Graal master ca436f5a8d260027da4218761ee1b83a4d0b4a0f
Quarkus version or git rev: Quarkus master dac8d8b
Build tool (ie. output of mvnw --version or gradlew --version): Apache Maven 3.6.3 (cecedd343002696d0abb50b32b541b8a6ba2883f)

The text was updated successfully, but these errors were encountered:

jerboaa · 2021-03-18T10:23:30Z

For reference, instead of building Graal VM master from source, one could also use 21.1 dev builds from:
https://github.com/graalvm/graalvm-ce-dev-builds/releases in order to reproduce those failures.

jerboaa · 2021-03-18T12:46:27Z

PR for solving the runtime-init problem for modules using http-client (includes integration-test/main): #15856

jerboaa · 2021-03-18T18:39:21Z

PR for integration-tests/redis-client: #15864

jerboaa · 2021-03-19T10:44:03Z

PR for integration-tests/reactive-messaging-kafka: #15879

galderz · 2021-03-19T13:24:11Z

I've created #15889 to track the quarkus-integration-test-bouncycastle-fips failures. It needs some consideration and agreement on the solution.

jerboaa · 2021-03-19T14:51:47Z

Elytron Security LDAP is actually a Graal issue:
oracle/graal#3297

jerboaa · 2021-03-19T18:47:07Z

PR for integration-tests/amazon-services: #15897

jerboaa · 2021-03-23T09:50:08Z

PR for integration-tests/artemis-core and integration-tests/artemis-jms: #15952

jerboaa · 2021-03-23T16:45:49Z

At this point most of the Random/SplittableRandom related failures should be fixed. I haven't done a full run, but it would be good to get one done with quarkus master to see what's still left. @zakkak could you help with that?

zakkak · 2021-03-23T22:03:20Z

@zakkak could you help with that?

Sure, I started https://github.com/zakkak/graalvm-quarkus-ci/actions/runs/681073992 to test latest Quarkus with latest GraalVM.

jerboaa · 2021-03-24T10:07:41Z

@zakkak Thanks. In the spirit of teach a man to fish, what would be the steps to do this myself? Anyway, most issues are now #15962 which is unfortunate since we still have lots of noise :(

zakkak · 2021-03-24T13:16:29Z

@zakkak Thanks. In the spirit of teach a man to fish, what would be the steps to do this myself?

In that case I just copied the GH workflow from oracle/graal to a new repo (zakkak/graalvm-quarkus-ci) and slightly edited it to use the latest dev branch of both graalvm and quarkus (the original workflow is using the latest release of quarkus with the latest dev branch of graalvm).

I also made it possible to trigger the workflow on demand with different combinations of versions and repos ~~as well. It would be nice to have the option to use forked repos as well~~ to try out PRs as well.

I probably need to give you (and anyone interested in running this kind of tests) some access rights to the repo.

Anyway, most issues are now #15962 which is unfortunate since we still have lots of noise :(

True, #15962 seems to be something new though ~~, so I'll try to schedule a run with an older version to see how far we have gone with the Random/SplittableRandom related failures.~~ I started a new workflow to test @geoand 's fix for #15962 with latest graalvm: https://github.com/zakkak/graalvm-quarkus-ci/actions/runs/683208450

jerboaa · 2021-03-24T16:38:43Z

@zakkak Thanks for the info!

zakkak · 2021-03-24T16:45:04Z

Job for latest Quarkus vs latest Graal now that #15962 is fixed:
https://github.com/zakkak/graalvm-quarkus-ci/actions/runs/683707199

geoand · 2021-03-24T17:17:11Z

👍🏼

jerboaa · 2021-03-24T17:57:25Z

Job for latest Quarkus vs latest Graal now that #15962 is fixed:
https://github.com/zakkak/graalvm-quarkus-ci/actions/runs/683707199

Now that looks a lot better already! Thank you.

zakkak · 2021-03-24T21:56:47Z

I list below the ITs that are still failing (based on https://github.com/zakkak/graalvm-quarkus-ci/actions/runs/683707199)

Random/SplitRandom related

Bouncy Castle FIPS
Elytron Security LDAP
JGit
Spring Data JPA
gRPC - TLS
gRPC - Plain Text with gzip compression
gRPC - Plain Text with Mutiny
gRPC - TLS with mutual auth
gRPC - Interceptors
gRPC - Proto v2

Other

Bouncy Castle JSEE
Compiles but fails with:

java.lang.NoSuchMethodException: org.bouncycastle.jcajce.provider.drbg.DRBG$Default.<init>(java.security.SecureRandomParameters)

Kafka with Snappy
Compiles but fails with:

 ERROR: Failed to start application (with profile prod)
org.xerial.snappy.SnappyError: [FAILED_TO_LOAD_NATIVE_LIBRARY] no native library is found for os.name=Linux and os.arch=x86_64
	at io.quarkus.kafka.client.runtime.KafkaRecorder.loadSnappy(KafkaRecorder.java:39)
	at io.quarkus.deployment.steps.KafkaProcessor$loadSnappyIfEnabled1534179476.deploy_0(KafkaProcessor$loadSnappyIfEnabled1534179476.zig:67)
	at io.quarkus.deployment.steps.KafkaProcessor$loadSnappyIfEnabled1534179476.deploy(KafkaProcessor$loadSnappyIfEnabled1534179476.zig:40)
	at io.quarkus.runner.ApplicationImpl.doStart(ApplicationImpl.zig:400)
	at io.quarkus.runtime.Application.start(Application.java:90)
	at io.quarkus.runtime.ApplicationLifecycleManager.run(ApplicationLifecycleManager.java:100)
	at io.quarkus.runtime.Quarkus.run(Quarkus.java:66)
	at io.quarkus.runtime.Quarkus.run(Quarkus.java:42)
	at io.quarkus.runtime.Quarkus.run(Quarkus.java:119)
	at io.quarkus.runner.GeneratedMain.main(GeneratedMain.zig:29)

zakkak · 2021-03-25T00:15:39Z

PR for gRPC* tests #16008

Backport of quarkus-qe#75. Won't run Native as it's not working on 1.11 due to quarkusio/quarkus#14904

Backport of #75. Won't run Native as it's not working on 1.11 due to quarkusio/quarkus#14904

akuzni2 · 2021-05-29T13:10:57Z

I'm getting this error when converting one of my applications to Quarkus to build a native-image using GraalVM 21.1.0. From this thread I can't really tell what I need to do to get around it.

I'm building my quarkus project with Quarkus 1.13.4.Final

galderz · 2021-06-14T09:59:20Z

@akuzni2 I might be wrong but I think 21.1 is only supported in Quarkus 2.x onwards...

zakkak · 2021-06-14T10:05:17Z

@akuzni2 I might be wrong but I think 21.1 is only supported in Quarkus 2.x onwards...

That's correct,

Quarkus version	Compatible GraalVM/Mandrel version
1.12.x	21.0.x.x
1.13.x	20.3.x.x 21.0.x.x
2.0.x	20.3.x.x 21.1.x.x

source: https://github.com/graalvm/mandrel/wiki

dufoli · 2021-06-24T08:26:36Z

hello,

I have same issue on cxf extension.
quarkiverse/quarkus-cxf#253
it failed with exception.
Can you help with that because I am not sure.
Do I need to switch to runtime initialization for some class :
org.apache.cxf.endpoint.ClientImpl

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. To see how this object got instantiated use --trace-object-instantiation=java.util.Random. The object was probably created by a class initializer and is reachable from a static field. You can request class initialization at image runtime by using the option --initialize-at-run-time=. Or you can write your own initialization methods and call them explicitly from your main entry point.
Detailed message:
Trace:
at parsing org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue(AttachmentUtil.java:255)
Call path from entry point to org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue():
at org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue(AttachmentUtil.java:255)
at org.apache.cxf.attachment.AttachmentSerializer.writeProlog(AttachmentSerializer.java:82)
at org.apache.cxf.interceptor.AttachmentOutInterceptor.handleMessage(AttachmentOutInterceptor.java:91)
at org.apache.cxf.phase.PhaseInterceptorChain.doIntercept(PhaseInterceptorChain.java:308)
at org.apache.cxf.endpoint.ClientImpl.onMessage(ClientImpl.java:831)
at org.apache.cxf.endpoint.ClientImpl$3$1.run(ClientImpl.java:923)
at java.lang.Thread.run(Thread.java:829)
at com.oracle.svm.core.thread.JavaThreads.threadStartRoutine(JavaThreads.java:553)
at com.oracle.svm.core.posix.thread.PosixJavaThreads.pthreadStartRoutine(PosixJavaThreads.java:192)
at com.oracle.svm.core.code.IsolateEnterStub.PosixJavaThreads_pthreadStartRoutine_e1f4a8c0039f8337338252cd8734f63a79b5e3df(generated:0)

com.oracle.svm.core.util.UserError$UserException: 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. To see how this object got instantiated use --trace-object-instantiation=java.util.Random. The object was probably created by a class initializer and is reachable from a static field. You can request class initialization at image runtime by using the option --initialize-at-run-time=. Or you can write your own initialization methods and call them explicitly from your main entry point.
Detailed message:
Trace:
at parsing org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue(AttachmentUtil.java:255)
Call path from entry point to org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue():
at org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue(AttachmentUtil.java:255)
at org.apache.cxf.attachment.AttachmentSerializer.writeProlog(AttachmentSerializer.java:82)
at org.apache.cxf.interceptor.AttachmentOutInterceptor.handleMessage(AttachmentOutInterceptor.java:91)
at org.apache.cxf.phase.PhaseInterceptorChain.doIntercept(PhaseInterceptorChain.java:308)
at org.apache.cxf.endpoint.ClientImpl.onMessage(ClientImpl.java:831)
at org.apache.cxf.endpoint.ClientImpl$3$1.run(ClientImpl.java:923)
at java.lang.Thread.run(Thread.java:829)
at com.oracle.svm.core.thread.JavaThreads.threadStartRoutine(JavaThreads.java:553)
at com.oracle.svm.core.posix.thread.PosixJavaThreads.pthreadStartRoutine(PosixJavaThreads.java:192)
at com.oracle.svm.core.code.IsolateEnterStub.PosixJavaThreads_pthreadStartRoutine_e1f4a8c0039f8337338252cd8734f63a79b5e3df(generated:0)

at com.oracle.svm.core.util.UserError.abort(UserError.java:82)
at com.oracle.svm.hosted.FallbackFeature.reportAsFallback(FallbackFeature.java:233)
at com.oracle.svm.hosted.NativeImageGenerator.runPointsToAnalysis(NativeImageGenerator.java:798)
at com.oracle.svm.hosted.NativeImageGenerator.doRun(NativeImageGenerator.java:582)
at com.oracle.svm.hosted.NativeImageGenerator.lambda$run$2(NativeImageGenerator.java:495)
at java.base/java.util.concurrent.ForkJoinTask$AdaptedRunnableAction.exec(ForkJoinTask.java:1407)
at java.base/java.util.concurrent.ForkJoinTask.doExec(ForkJoinTask.java:290)
at java.base/java.util.concurrent.ForkJoinPool$WorkQueue.topLevelExec(ForkJoinPool.java:1020)
at java.base/java.util.concurrent.ForkJoinPool.scan(ForkJoinPool.java:1656)
at java.base/java.util.concurrent.ForkJoinPool.runWorker(ForkJoinPool.java:1594)
at java.base/java.util.concurrent.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:183)
Caused by: com.oracle.graal.pointsto.constraints.UnsupportedFeatureException: 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. To see how this object got instantiated use --trace-object-instantiation=java.util.Random. The object was probably created by a class initializer and is reachable from a static field. You can request class initialization at image runtime by using the option --initialize-at-run-time=. Or you can write your own initialization methods and call them explicitly from your main entry point.
Detailed message:
Trace:
at parsing org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue(AttachmentUtil.java:255)
Call path from entry point to org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue():
at org.apache.cxf.attachment.AttachmentUtil.getUniqueBoundaryValue(AttachmentUtil.java:255)
at org.apache.cxf.attachment.AttachmentSerializer.writeProlog(AttachmentSerializer.java:82)
at org.apache.cxf.interceptor.AttachmentOutInterceptor.handleMessage(AttachmentOutInterceptor.java:91)
at org.apache.cxf.phase.PhaseInterceptorChain.doIntercept(PhaseInterceptorChain.java:308)
at org.apache.cxf.endpoint.ClientImpl.onMessage(ClientImpl.java:831)
at org.apache.cxf.endpoint.ClientImpl$3$1.run(ClientImpl.java:923)
at java.lang.Thread.run(Thread.java:829)
at com.oracle.svm.core.thread.JavaThreads.threadStartRoutine(JavaThreads.java:553)
at com.oracle.svm.core.posix.thread.PosixJavaThreads.pthreadStartRoutine(PosixJavaThreads.java:192)
at com.oracle.svm.core.code.IsolateEnterStub.PosixJavaThreads_pthreadStartRoutine_e1f4a8c0039f8337338252cd8734f63a79b5e3df(generated:0)

at com.oracle.graal.pointsto.constraints.UnsupportedFeatures.report(UnsupportedFeatures.java:126)
at com.oracle.svm.hosted.NativeImageGenerator.runPointsToAnalysis(NativeImageGenerator.java:795)
... 8 more
Caused by: 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. To see how this object got instantiated use --trace-object-instantiation=java.util.Random. The object was probably created by a class initializer and is reachable from a static field. You can request class initialization at image runtime by using the option --initialize-at-run-time=. Or you can write your own initialization methods and call them explicitly from your main entry point.
at com.oracle.svm.hosted.image.DisallowedImageHeapObjectFeature.error(DisallowedImageHeapObjectFeature.java:163)
at com.oracle.svm.core.image.DisallowedImageHeapObjects.check(DisallowedImageHeapObjects.java:65)
at com.oracle.svm.hosted.image.DisallowedImageHeapObjectFeature.replacer(DisallowedImageHeapObjectFeature.java:139)
at com.oracle.graal.pointsto.meta.AnalysisUniverse.replaceObject(AnalysisUniverse.java:565)
at com.oracle.svm.hosted.ameta.AnalysisConstantReflectionProvider.replaceObject(AnalysisConstantReflectionProvider.java:213)
at com.oracle.svm.hosted.ameta.AnalysisConstantReflectionProvider.interceptValue(AnalysisConstantReflectionProvider.java:184)
at com.oracle.svm.hosted.ameta.AnalysisConstantReflectionProvider.readValue(AnalysisConstantReflectionProvider.java:98)
at com.oracle.svm.hosted.ameta.AnalysisConstantReflectionProvider.readFieldValue(AnalysisConstantReflectionProvider.java:77)
at jdk.internal.vm.compiler/org.graalvm.compiler.nodes.util.ConstantFoldUtil$1.readValue(ConstantFoldUtil.java:51)
at jdk.internal.vm.compiler/org.graalvm.compiler.core.common.spi.JavaConstantFieldProvider.readConstantField(JavaConstantFieldProvider.java:84)
at com.oracle.svm.hosted.ameta.AnalysisConstantFieldProvider.readConstantField(AnalysisConstantFieldProvider.java:70)
at jdk.internal.vm.compiler/org.graalvm.compiler.nodes.util.ConstantFoldUtil.tryConstantFold(ConstantFoldUtil.java:47)
at com.oracle.svm.hosted.phases.ConstantFoldLoadFieldPlugin.tryConstantFold(ConstantFoldLoadFieldPlugin.java:61)
at com.oracle.svm.hosted.phases.ConstantFoldLoadFieldPlugin.handleLoadStaticField(ConstantFoldLoadFieldPlugin.java:57)
at jdk.internal.vm.compiler/org.graalvm.compiler.java.BytecodeParser.genGetStatic(BytecodeParser.java:4972)
at jdk.internal.vm.compiler/org.graalvm.compiler.java.BytecodeParser.genGetStatic(BytecodeParser.java:4939)
at jdk.internal.vm.compiler/org.graalvm.compiler.java.BytecodeParser.processBytecode(BytecodeParser.java:5442)
at jdk.internal.vm.compiler/org.graalvm.compiler.java.BytecodeParser.iterateBytecodesForBlock(BytecodeParser.java:3451)
at jdk.internal.vm.compiler/org.graalvm.compiler.java.BytecodeParser.processBlock(BytecodeParser.java:3258)
at jdk.internal.vm.compiler/org.graalvm.compiler.java.BytecodeParser.build(BytecodeParser.java:1125)
at jdk.internal.vm.compiler/org.graalvm.compiler.java.BytecodeParser.buildRootMethod(BytecodeParser.java:1019)
at jdk.internal.vm.compiler/org.graalvm.compiler.java.GraphBuilderPhase$Instance.run(GraphBuilderPhase.java:84)
at com.oracle.svm.hosted.phases.SharedGraphBuilderPhase.run(SharedGraphBuilderPhase.java:76)
at jdk.internal.vm.compiler/org.graalvm.compiler.phases.Phase.run(Phase.java:49)
at jdk.internal.vm.compiler/org.graalvm.compiler.phases.BasePhase.apply(BasePhase.java:212)
at jdk.internal.vm.compiler/org.graalvm.compiler.phases.Phase.apply(Phase.java:42)
at jdk.internal.vm.compiler/org.graalvm.compiler.phases.Phase.apply(Phase.java:38)
at com.oracle.graal.pointsto.flow.AnalysisParsedGraph.parseBytecode(AnalysisParsedGraph.java:113)
at com.oracle.svm.hosted.SVMHost.parseBytecode(SVMHost.java:647)
at com.oracle.graal.pointsto.meta.AnalysisMethod.ensureGraphParsed(AnalysisMethod.java:592)
at com.oracle.graal.pointsto.flow.MethodTypeFlowBuilder.parse(MethodTypeFlowBuilder.java:163)
at com.oracle.graal.pointsto.flow.MethodTypeFlowBuilder.apply(MethodTypeFlowBuilder.java:304)
at com.oracle.graal.pointsto.flow.MethodTypeFlow.createTypeFlow(MethodTypeFlow.java:313)
at com.oracle.graal.pointsto.flow.MethodTypeFlow.ensureTypeFlowCreated(MethodTypeFlow.java:302)
at com.oracle.graal.pointsto.flow.MethodTypeFlow.addContext(MethodTypeFlow.java:103)
at com.oracle.graal.pointsto.flow.StaticInvokeTypeFlow.update(InvokeTypeFlow.java:420)
at com.oracle.graal.pointsto.BigBang$2.run(BigBang.java:547)
at com.oracle.graal.pointsto.util.CompletionExecutor.lambda$execute$0(CompletionExecutor.java:173)
at java.base/java.util.concurrent.ForkJoinTask$RunnableExecuteAction.exec(ForkJoinTask.java:1426)

zakkak · 2021-06-28T08:52:38Z

Hi @dufoli, my understanding is that quarkiverse/quarkus-cxf@8804dda fixed the above issue, is that right?

dufoli · 2021-06-28T09:08:02Z

@zakkak yes

akuzni2 · 2021-06-28T22:57:49Z

HI @zakkak @dufoli thanks for commenting back. I've upgraded to Quarkus 2.0.0 and I still see the same issue cause by some AWS dependencies.

Error: 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.  To see how this object got instantiated use --trace-object-instantiation=java.util.Random. The object was probably created by a class initializer and is reachable from a static field. You can request class initialization at image runtime by using the option --initialize-at-run-time=<class-name>. Or you can write your own initialization methods and call them explicitly from your main entry point.
Trace: Object was reached by 
        reading field com.amazonaws.retry.PredefinedBackoffStrategies$EqualJitterBackoffStrategy.random of
                constant com.amazonaws.retry.PredefinedBackoffStrategies$EqualJitterBackoffStrategy@546fe072 reached by 
        reading field com.amazonaws.retry.PredefinedBackoffStrategies$SDKDefaultBackoffStrategy.equalJitterBackoffStrategy of
                constant com.amazonaws.retry.PredefinedBackoffStrategies$SDKDefaultBackoffStrategy@4d11ec9e reached by 
        reading field com.amazonaws.retry.RetryPolicy.backoffStrategy of
                constant com.amazonaws.retry.RetryPolicy@28b37dae reached by 
        scanning method com.amazonaws.ClientConfiguration.<init>(ClientConfiguration.java:155)
Call path from entry point to com.amazonaws.ClientConfiguration.<init>(ClientConfiguration): 
        at com.amazonaws.ClientConfiguration.<init>(ClientConfiguration.java:378)
        at com.amazonaws.client.builder.AwsClientBuilder.resolveClientConfiguration(AwsClientBuilder.java:169)
        at com.amazonaws.client.builder.AwsClientBuilder.access$000(AwsClientBuilder.java:54)
        at com.amazonaws.client.builder.AwsClientBuilder$SyncBuilderParams.<init>(AwsClientBuilder.java:505)
        at com.amazonaws.client.builder.AwsClientBuilder.getSyncClientParams(AwsClientBuilder.java:441)
        at com.amazonaws.client.builder.AwsSyncClientBuilder.build(AwsSyncClientBuilder.java:46)
        at com.mycompany.service.ApiServiceImpl.<init>(ApiServiceImpl.java:23)
        at com.mycompany.service.ProcessingService.<init>(ProcessingService.java:22)
        at com.mycompany.service.ProcessingService_Bean.create(ProcessingService_Bean.zig:140)
        at com.mycompany.service.ProcessingService_Bean.create(ProcessingService_Bean.zig:155)
        at io.quarkus.arc.impl.AbstractSharedContext.createInstanceHandle(AbstractSharedContext.java:96)
        at io.quarkus.arc.impl.AbstractSharedContext$1.get(AbstractSharedContext.java:29)
        at io.quarkus.arc.impl.AbstractSharedContext$1.get(AbstractSharedContext.java:26)
        at com.oracle.svm.core.jdk.SystemPropertiesSupport.initializeLazyValue(SystemPropertiesSupport.java:216)
        at com.oracle.svm.core.jdk.SystemPropertiesSupport.getProperty(SystemPropertiesSupport.java:169)
        at com.oracle.svm.core.jdk.Target_java_lang_System.getProperty(JavaLangSubstitutions.java:291)
        at com.oracle.svm.jni.JNIJavaCallWrappers.jniInvoke_VA_LIST:Ljava_lang_System_2_0002egetProperty_00028Ljava_lang_String_2_00029Ljava_lang_String_2(generated:0)

I'm not really sure how to perform the recommendation by the error message

Or you can write your own initialization methods and call them explicitly from your main entry point.

If you have any ideas please let me know.

dufoli · 2021-06-29T06:35:15Z

@akuzni2 what aws extension are you using ?
do you add the corresponding quarkus extension ?
s3, lambda, SNS, SQS, ... ?

wsy89 · 2021-07-05T05:26:29Z

@dufoli I am getting exactly the same issue as @akuzni2 is getting when I try to invoke another lamba using

@Produces
@ApplicationScoped
public AWSLambda awsLambda() {
    return AWSLambdaAsyncClient.builder().build();
}

<dependency>
    <groupId>com.amazonaws</groupId>
    <artifactId>aws-java-sdk-lambda</artifactId>
    <version>1.12.15</version>
</dependency>

EdwinKorsten · 2021-07-05T15:37:34Z

I have the same problem as @wsy89 and @akuzni2 with an AWS SDK library.

Error: 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. To see how this object got instantiated use --trace-object-instantiation=java.util.Random. The object was probably created by a class initializer and is reachable from a static field. You can request class initialization at image runtime by using the option --initialize-at-run-time=. Or you can write your own initialization methods and call them explicitly from your main entry point.
Trace: Object was reached by
reading field com.amazonaws.retry.PredefinedBackoffStrategies$EqualJitterBackoffStrategy.random of
constant com.amazonaws.retry.PredefinedBackoffStrategies$EqualJitterBackoffStrategy@5c3fa02c reached by
reading field com.amazonaws.retry.PredefinedBackoffStrategies$SDKDefaultBackoffStrategy.equalJitterBackoffStrategy of
constant com.amazonaws.retry.PredefinedBackoffStrategies$SDKDefaultBackoffStrategy@42c1d66e reached by
reading field com.amazonaws.retry.RetryPolicy.backoffStrategy of
constant com.amazonaws.retry.RetryPolicy@35cbe86 reached by
scanning method com.amazonaws.ClientConfiguration.(ClientConfiguration.java:74)
Call path from entry point to com.amazonaws.ClientConfiguration.():
no path found from entry point to target method

Adding this to maven does not help:
<quarkus.native.additional-build-args>--initialize-at-run-time=com.amazonaws.ClientConfiguration,com.amazonaws.retry.PredefinedBackoffStrategies</quarkus.native.additional-build-args>

wsy89 · 2021-07-05T22:56:58Z

@EdwinKorsten @akuzni2
I was able to fix this by adding

--initialize-at-run-time=org.apache.http.impl.auth.NTLMEngineImpl

in additional build arg.
Also I had to use aws sdk v2 instead of v1

antonwiens · 2021-07-07T16:19:51Z

@gsmet I am running also into the issue with org.apache.http.impl.auth.NTLMEngineImpl using aws s3 extension and apache client. Can we get a fix for that?

jerboaa · 2021-07-07T16:31:39Z

@antonwiens Are you also using the HTTP client extension as a dependency as described here?

https://quarkus.io/guides/amazon-s3#configuring-s3-clients

antonwiens · 2021-07-07T16:35:17Z

@jerboaa No, i didn't. I pulled up the versions and didn`t know i had to add this, thats a nice hint, thanks and sorry for pulling this issue up again.

Maybe a build error from the quarkus plugin would be nice, if the apache client dependency was added and the quarkus-apache-client extension was missing. Just as a hint for ppl upgrading or not reading the documentation properly :)

jmgtan · 2022-03-03T08:31:56Z

I've been getting the same Random/SplittableRandom error. I'm currently following the Quarkus Lambda guide (https://quarkus.io/guides/amazon-lambda) using the Quarkus 2.7.3.Final with GraalVM 21.3.1. I'm using the DynamoDB client from Java SDK v2.

Update: managed to get it to work by following the sample code found here: https://github.com/quarkusio/quarkus-quickstarts/blob/main/amazon-dynamodb-quickstart/pom.xml. Specifically by the use of the quarkus-amazon-dynamodb dependency instead of including the AWS SDK directly. I don't see this mentioned in any of the guides. Working now and man is the Lambda function fast! Great work.

galderz · 2022-03-04T07:56:07Z

@jmgtan Maybe you could improve the amazon-lambda guide? OSS FTW

MarkusKramer · 2022-07-25T13:32:30Z

@jmgtan I've also started hitting the Random/SplittableRandom error with the latest AWS SDK. It is possible to workaround it by using --initialize-at-run-time=software.amazon.awssdk.utils.cache.CachedSupplier. see aws/aws-sdk-java-v2#3323

gsmet · 2022-07-25T14:41:48Z

@MarkusKramer I created quarkiverse/quarkus-amazon-services#220 to fix the issue in our extension.

zakkak added the kind/bug Something isn't working label Feb 8, 2021

ghost added the triage/needs-triage label Feb 8, 2021

zakkak added area/native-image and removed triage/needs-triage labels Feb 8, 2021

zakkak mentioned this issue Feb 8, 2021

Fix Flyway extension in relation with GraalVM master #14908

Merged

zakkak mentioned this issue Mar 11, 2021

Quarkus Nightly consistently failing oracle/graal#3272

Closed

jerboaa mentioned this issue Mar 18, 2021

Initialize resteasy dependency class using SecureRandom at runtime #15846

Closed

jerboaa mentioned this issue Mar 19, 2021

Initialize classes using j.u.Random at runtime in kafka-client #15879

Merged

galderz mentioned this issue Mar 19, 2021

Bouncycastle FIPs leaving a SecureRandom in heap #15889

Closed

jerboaa mentioned this issue Mar 19, 2021

Runtime initialize amazon-services classes using j.u.Random #15897

Merged

jerboaa mentioned this issue Mar 23, 2021

Runtime initialize RandomUtil in artemis-core extension #15952

Merged

zakkak mentioned this issue Mar 25, 2021

Runtime initialize grpc class using j.u.Random #16008

Merged

Sgitario added a commit to Sgitario/quarkus-startstop that referenced this issue May 5, 2021

Adapt Start Stop TS to Windows

23f77ba

Backport of quarkus-qe#75. Won't run Native as it's not working on 1.11 due to quarkusio/quarkus#14904

Sgitario mentioned this issue May 5, 2021

[1.11] Adapt Start Stop TS to Windows quarkus-qe/quarkus-startstop#79

Merged

Sgitario added a commit to Sgitario/quarkus-startstop that referenced this issue May 10, 2021

Adapt Start Stop TS to Windows

1493b26

Backport of quarkus-qe#75. Won't run Native as it's not working on 1.11 due to quarkusio/quarkus#14904

rsvoboda pushed a commit to quarkus-qe/quarkus-startstop that referenced this issue May 11, 2021

Adapt Start Stop TS to Windows

0f9c3bb

Backport of #75. Won't run Native as it's not working on 1.11 due to quarkusio/quarkus#14904

pjgg mentioned this issue May 14, 2021

Building native images with quarkus-opentelemetry extension throw a Random/SplittableRandom initialized error at build time #17234

Closed

nastra mentioned this issue Jun 24, 2021

Upgrade to Quarkus 2.0.0.Final projectnessie/nessie#1490

Merged

dufoli mentioned this issue Jun 24, 2021

Bump quarkus.version from 1.13.7.Final to 2.0.0.Final quarkiverse/quarkus-cxf#253

Merged

ffang mentioned this issue Jul 15, 2021

can't build camel-fhir extension native image apache/camel-quarkus#2906

Closed

jerboaa mentioned this issue Mar 3, 2022

Native-Image: ApplicationImpl Initialize on runtime when using JAX-RS-Filter #23716

Open

MarkusKramer mentioned this issue Jul 26, 2022

Native image build fails due to CachedSupplier/SplittableRandom aws/aws-sdk-java-v2#3323

Closed

michalvavrik mentioned this issue Aug 8, 2022

Update native image builders - use version 22.2 quarkus-qe/quarkus-extensions-combinations#155

Merged

ppalaga mentioned this issue Sep 26, 2022

Instance of Random/SplittableRandom class in CachedSupplier.jitterTime apache/camel-quarkus#4134

Open

Native image build failures on latest GraalVM with Random/SplittableRandom initialized at build time #14904

Native image build failures on latest GraalVM with Random/SplittableRandom initialized at build time #14904

Comments

zakkak commented Feb 8, 2021 • edited Loading

jerboaa commented Mar 18, 2021

jerboaa commented Mar 18, 2021 • edited Loading

jerboaa commented Mar 18, 2021

jerboaa commented Mar 19, 2021

galderz commented Mar 19, 2021

jerboaa commented Mar 19, 2021

jerboaa commented Mar 19, 2021

jerboaa commented Mar 23, 2021

jerboaa commented Mar 23, 2021

zakkak commented Mar 23, 2021

jerboaa commented Mar 24, 2021

zakkak commented Mar 24, 2021 • edited Loading

jerboaa commented Mar 24, 2021

zakkak commented Mar 24, 2021

geoand commented Mar 24, 2021

jerboaa commented Mar 24, 2021

zakkak commented Mar 24, 2021

Random/SplitRandom related

Other

zakkak commented Mar 25, 2021 • edited Loading

akuzni2 commented May 29, 2021

galderz commented Jun 14, 2021

zakkak commented Jun 14, 2021

dufoli commented Jun 24, 2021

zakkak commented Jun 28, 2021 • edited Loading

dufoli commented Jun 28, 2021

akuzni2 commented Jun 28, 2021

dufoli commented Jun 29, 2021

wsy89 commented Jul 5, 2021

EdwinKorsten commented Jul 5, 2021

wsy89 commented Jul 5, 2021

antonwiens commented Jul 7, 2021

jerboaa commented Jul 7, 2021

antonwiens commented Jul 7, 2021

jmgtan commented Mar 3, 2022 • edited Loading

galderz commented Mar 4, 2022

MarkusKramer commented Jul 25, 2022

gsmet commented Jul 25, 2022

zakkak commented Feb 8, 2021 •

edited

Loading

jerboaa commented Mar 18, 2021 •

edited

Loading

zakkak commented Mar 24, 2021 •

edited

Loading

zakkak commented Mar 25, 2021 •

edited

Loading

zakkak commented Jun 28, 2021 •

edited

Loading

jmgtan commented Mar 3, 2022 •

edited

Loading