dd-trace-java v1.31.2 crashes the JVM

[henrik-sandberg]

Hi DD team,

We've experienced JVM crashes due to segmentation fault with references to dd profiling. It's a containerized application running versions.

Tracer version: v1.31.2 Java VM: OpenJDK 64-Bit Server VM Temurin-21.0.2+13 OS: Alpine Linux v3.19 Datadog agent: 7.42.1

The JVM have mostly crashed within the first two minutes after start, but has on more rare occasions crashed up to 6 hours after application start.

I will ensure to bring versions up to speed, and meanwhile we'll disable the profiling with DD_PROFILING_ENABLED=false.

Attached are two log entries.

# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007f82be65e43b, pid=1, tid=31
#
# JRE version: OpenJDK Runtime Environment Temurin-21.0.2+13 (21.0.2+13) (build 21.0.2+13-LTS)
# Java VM: OpenJDK 64-Bit Server VM Temurin-21.0.2+13 (21.0.2+13-LTS, mixed mode, sharing, tiered, compressed oops, compressed class ptrs, g1 gc, linux-amd64)
# Problematic frame:
# C  [ld-musl-x86_64.so.1+0x5943b]  strlen+0xd
#
# Core dump will be written. Default location: //core.1
#
# The JFR repository may contain useful JFR files. Location: /tmp/2024_04_10_09_23_20_1
#
# If you would like to submit a bug report, please visit:
#   https://github.com/adoptium/adoptium-support/issues
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#
---------------  S U M M A R Y ------------
Command Line: -javaagent:./agents/dd-java-agent.jar -XX:FlightRecorderOptions=stackdepth=256 -Dcom.sun.org.apache.xml.internal.security.ignoreLineBreaks=true -Dcom.ibm.mq.cfg.useIBMCipherMappings=false /app.jar
Host: Intel(R) Xeon(R) CPU E5-2697A v4 @ 2.60GHz, 2 cores, 3G, Alpine Linux v3.19
Time: Wed Apr 10 09:24:21 2024 CEST elapsed time: 127.619058 seconds (0d 0h 2m 7s)
---------------  T H R E A D  ---------------
Current thread (0x00007f825afce880):  JavaThread “dd-profiler-recording-scheduler” daemon [_thread_in_native, id=31, stack(0x00007f825ad2f000,0x00007f825ae2fa78) (1026K)]
Stack: [0x00007f825ad2f000,0x00007f825ae2fa78],  sp=0x00007f825ae2eb48,  free space=1022k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
C  [ld-musl-x86_64.so.1+0x5943b]  strlen+0xd
C  [libjavaProfiler2790011097090240288.so+0x18737]  Recording::writeSettings(Buffer*, Arguments&)+0xd7
C  [libjavaProfiler2790011097090240288.so+0x25b09]  Recording::switchChunk(int)+0xf9
C  [libjavaProfiler2790011097090240288.so+0x25bef]  FlightRecorder::dump(char const*, int)+0x6f
C  [libjavaProfiler2790011097090240288.so+0x3cfd1]  Profiler::dump(char const*, int)+0x201
C  [libjavaProfiler2790011097090240288.so+0x48f37]  Java_com_datadoghq_profiler_JavaProfiler_dump0+0x47

Event: 11.935 Thread 0x00007f8174166820 Exception <a ‘java/lang/IncompatibleClassChangeError’{0x00000000c71b2228}: Found class java.lang.Object, but interface was expected> (0x00000000c71b2228)
thrown [src/hotspot/share/interpreter/linkResolver.cpp, line 840]
Event: 11.978 Thread 0x00007f81131017c0 Exception <a ‘java/lang/NoSuchMethodError’{0x00000000c944ba80}: ‘long java.lang.invoke.DirectMethodHandle$Holder.invokeStatic(java.lang.Object, java.lang.Object, java.lang.Object)’> (0x00000000c944ba80)
thrown [src/hotspot/share/interpreter/linkResolver.cpp, line 772]
Event: 12.024 Thread 0x00007f81131017c0 Exception <a ‘java/lang/NoSuchMethodError’{0x00000000c91ecb48}: ‘java.lang.Object java.lang.invoke.DirectMethodHandle$Holder.newInvokeSpecial(java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, int)’> (0x00000000c91ecb48)
thrown [src/hotspot/share/interpreter/linkResolver.cpp, line 772]
Event: 12.025 Thread 0x00007f81131017c0 Exception <a ‘java/lang/NoSuchMethodError’{0x00000000c9083eb0}: ‘java.lang.Object java.lang.invoke.DirectMethodHandle$Holder.invokeSpecial(java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, java.lang.Object, int)’> (0x00000000c9083eb0)
thrown [src/hotspot/share/interpreter/linkResolver.cpp, line 772]
Event: 14.664 Thread 0x00007f810f692790 Exception <a ‘java/lang/NoSuchMethodError’{0x00000000c7c99770}: ‘double java.lang.invoke.DelegatingMethodHandle$Holder.reinvoke_L(java.lang.Object, java.lang.Object)’> (0x00000000c7c99770)
thrown [src/hotspot/share/interpreter/linkResolver.cpp, line 772]
Event: 14.666 Thread 0x00007f810f692790 Exception <a ‘java/lang/NoSuchMethodError’{0x00000000c7cbe398}: ‘double java.lang.invoke.DirectMethodHandle$Holder.invokeSpecial(java.lang.Object, java.lang.Object)’> (0x00000000c7cbe398)
thrown [src/hotspot/share/interpreter/linkResolver.cpp, line 772]
Event: 14.851 Thread 0x00007f81130310a0 Exception <a ‘java/lang/NoSuchMethodError’{0x00000000c79687e0}: ‘java.lang.Object java.lang.invoke.DirectMethodHandle$Holder.invokeStatic(java.lang.Object, double, java.lang.Object)’> (0x00000000c79687e0)
thrown [src/hotspot/share/interpreter/linkResolver.cpp, line 772]
Event: 14.852 Thread 0x00007f81130310a0 Exception <a ‘java/lang/NoSuchMethodError’{0x00000000c796e298}: ‘java.lang.Object java.lang.invoke.DirectMethodHandle$Holder.newInvokeSpecial(java.lang.Object, double)’> (0x00000000c796e298)
[11:22 AM] ==============
dd-profiler-recording-scheduler
[11:23 AM] #
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007f817573343b, pid=1, tid=32
#
# JRE version: OpenJDK Runtime Environment Temurin-21.0.2+13 (21.0.2+13) (build 21.0.2+13-LTS)
# Java VM: OpenJDK 64-Bit Server VM Temurin-21.0.2+13 (21.0.2+13-LTS, mixed mode, sharing, tiered, compressed oops, compressed class ptrs, g1 gc, linux-amd64)
# Problematic frame:
# C  [ld-musl-x86_64.so.1+0x5943b]  strlen+0xd
#
# Core dump will be written. Default location: //core.1
#
# The JFR repository may contain useful JFR files. Location: /tmp/2024_04_10_10_37_03_1
#
# If you would like to submit a bug report, please visit:
#   https://github.com/adoptium/adoptium-support/issues
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#
---------------  S U M M A R Y ------------
Command Line: -javaagent:./agents/dd-java-agent.jar -XX:FlightRecorderOptions=stackdepth=256 -Dcom.sun.org.apache.xml.internal.security.ignoreLineBreaks=true -Dcom.ibm.mq.cfg.useIBMCipherMappings=false /app.jar
Host: Intel(R) Xeon(R) CPU E5-2697A v4 @ 2.60GHz, 2 cores, 3G, Alpine Linux v3.19
Time: Wed Apr 10 10:38:04 2024 CEST elapsed time: 118.902752 seconds (0d 0h 1m 58s)
---------------  T H R E A D  ---------------
Current thread (0x00007f8113100040):  JavaThread “dd-profiler-recording-scheduler” daemon [_thread_in_native, id=32, stack(0x00007f811226a000,0x00007f811236aa78) (1026K)]
Stack: [0x00007f811226a000,0x00007f811236aa78],  sp=0x00007f8112369af8,  free space=1022k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
C  [ld-musl-x86_64.so.1+0x5943b]  strlen+0xd
C  [libjavaProfiler1637923034457328543.so+0x18737]  Recording::writeSettings(Buffer*, Arguments&)+0xd7
C  [libjavaProfiler1637923034457328543.so+0x25b09]  Recording::switchChunk(int)+0xf9
C  [libjavaProfiler1637923034457328543.so+0x25bef]  FlightRecorder::dump(char const*, int)+0x6f
C  [libjavaProfiler1637923034457328543.so+0x3cfd1]  Profiler::dump(char const*, int)+0x201
C  [libjavaProfiler1637923034457328543.so+0x48f37]  Java_com_datadoghq_profiler_JavaProfiler_dump0+0x47
j  com.datadoghq.profiler.JavaProfiler.dump0(Ljava/lang/String;)V+0
j  com.datadoghq.profiler.JavaProfiler.dump(Ljava/nio/file/Path;)V+11
j  com.datadog.profiling.ddprof.DatadogProfiler.dump(Ljava/nio/file/Path;)V+5
j  com.datadog.profiling.ddprof.DatadogProfilerRecording.snapshot(Ljava/time/Instant;Ldatadog/trace/api/profiling/ProfilingSnapshot$Kind;)Ldatadog/trace/api/profiling/RecordingData;+17
j  com.datadog.profiling.controller.ddprof.DatadogProfilerOngoingRecording.snapshot(Ljava/time/Instant;Ldatadog/trace/api/profiling/ProfilingSnapshot$Kind;)Ldatadog/trace/api/profiling/RecordingData;+10
j  com.datadog.profiling.agent.CompositeController$CompositeOngoingRecording.lambda$snapshot$0(Ljava/time/Instant;Ldatadog/trace/api/profiling/ProfilingSnapshot$Kind;Lcom/datadog/profiling/controller/OngoingRecording;)Ldatadog/trace/api/profiling/RecordingData;+3
j  com.datadog.profiling.agent.CompositeController$CompositeOngoingRecording$$Lambda+0x00007f811b937ac8.apply(Ljava/lang/Object;)Ljava/lang/Object;+12
J 2779 c1 java.util.stream.ReferencePipeline$3$1.accept(Ljava/lang/Object;)V [email protected] (23 bytes) @ 0x00007f815f48eedc [0x00007f815f48edc0+0x000000000000011c]
J 14957 c1 java.util.ArrayList$ArrayListSpliterator.forEachRemaining(Ljava/util/function/Consumer;)V [email protected] (127 bytes) @ 0x00007f815f2c768c [0x00007f815f2c7400+0x000000000000028c]
J 14388 c2 java.util.stream.ReferencePipeline.collect(Ljava/util/stream/Collector;)Ljava/lang/Object; [email protected] (124 bytes) @ 0x00007f81672162bc [0x00007f8167215fe0+0x00000000000002dc]
j  com.datadog.profiling.agent.CompositeController$CompositeOngoingRecording.compose(Ljava/util/function/Function;)Ldatadog/trace/api/profiling/RecordingData;+22
j  com.datadog.profiling.agent.CompositeController$CompositeOngoingRecording.snapshot(Ljava/time/Instant;Ldatadog/trace/api/profiling/ProfilingSnapshot$Kind;)Ldatadog/trace/api/profiling/RecordingData;+8
j  com.datadog.profiling.controller.ProfilingSystem$SnapshotRecording.snapshot(Z)V+38

[jbachorik]

Hi Henrik, my apologies for inconvenience.

First, you don't need to disable profiling completely - DD_PROFILING_DDPROF_ENABLED=false will disable the profiling library where the crash occurs, still keeping the JFR profiling and most of the profiling functionality active.

I wonder if you could attach the full hs_err.log file here. The code is heavily inlined and having all information from the crash log would help us to diagnose the root cause.

Thanks for understanding.

[henrik-sandberg]

Hi Henrik, my apologies for inconvenience.

First, you don't need to disable profiling completely - DD_PROFILING_DDPROF_ENABLED=false will disable the profiling library where the crash occurs, still keeping the JFR profiling and most of the profiling functionality active.

I wonder if you could attach the full hs_err.log file here. The code is heavily inlined and having all information from the crash log would help us to diagnose the root cause.

Thanks for understanding.

Hi @jbachorik ,

Thanks for the quick reply! I have a log file ready on this error, however I would prefer a closed communication channel given I can't affirm the information it contains . Do you have any alternative ways I can share the file?

[jbachorik]

HI @henrik-sandberg - feel free to open a support ticket and then give me the ticket number. I should be able to follow up.

[henrik-sandberg]

Thank you @jbachorik - I have created #1645495. Please let me know if I can assist further in any way.