EPIC: jdk_jfr issues across multiple platforms

[smlambert]

This is an umbrella issue under which to track the large number of jfr issues seen across multiple platforms.

This issue will remain open and used to exclude tests under, until issues can be fixed (product and/or test updates upstream, etc), and allow us to close the multiple issues for specific versions and platforms that have accumulated in this repository since we have been running the jdk_jfr target.

• jdk/jfr/event/oldobject/TestObjectDescription.java timeout by jdk8u #4342
• jdk18: jdk/jfr/event/runtime/TestDirectBufferStatisticsEvent.java intermittent failure on aarch64_linux #3853
• jdk11 win32: jdk/jfr/api/recording/state/TestStateInvalid.java got [error][jfr] Could not finish chunk #3836
• jdk17+ : jdk/jfr/event/gc/detailed/TestGCLockerEvent.java timeout on windows #3817
• TestPromotionFailedEventWithParNew.java intermittently fail on linux-arch64 by temurin8 #3770
• TestThrowableInstrumentation.java intermittently fail by temurin8 on linux-aarch64 #3760
• jdk8 run TestCompilerInlining.java intermittently fail #3730
• TestG1OldAllocationPendingStackTrace.java run timeout because of can't trigger GC on some machine #3709
• jdk/jfr/jcmd/TestJcmdDump.java.TestJcmdDump consistently failing in jdk18+36 Release Triage #3492
• jdk_jfr_2 fail on jdk8 arm32 linux #3301
• jdk_jfr tests fail on jdk17 windows32 #3277
• jdk_jfr tests fail on jdk8 hotspot xLinux #3253
• jdk/jfr/event/io/TestInstrumentation fails on JDK8 Mac Hotspot #3249
• jdk_jfr tests fail on ppc linux jdk8 hotspot #3247
• Multiple tests from jdk_jfr_2 fail on arm32 due to dump related issues #3115
• JFR Test TestIsEnabledMultiple fails on Windows 32 for Java 11 (only failed on 11.0.13 release) #3053
• JFR TestCompilerCompile times out on Windows 32 #3046
• jfr TestCompilerPhase test fails on win-32 #3045
• JFR TestCodeCacheFull crashes JVM on Mac #3042
• Exclude 2 testcases in jdk_jfr test target that FAILED in Test_openjdk8_hs_extended.openjdk_aarch64_linux #2985
• jdk_jfr FAILED in Test_openjdk8_hs_extended.openjdk_x86-64_solaris_testList_0 #2919
• jdk17+35: Win32: EXCEPTION_ACCESS_VIOLATION (0xc0000005) : # V [jvm.dll+0x29f34] #2872
• jdk16+ jdk/jfr/event/allocation/TestObjectAllocationInNewTLABEvent.java got ava.lang.RuntimeException: Too few tlab objects allocated: expected 50 >= 80 on win32 and arm-linux  #2870
• jdk16+ 6 JFR tests got timeout on win32 and arm-linux #2866
• jdk16+ JFR test jdk/jfr/jmx/streaming/TestClose.java got java.lang.Exception: Expected repository to be empty on win32 and arm-linux #2865
• JDK16+ some JFR tests failed with fatal errors on window32 #2855
• jdk8: jdk/jfr/jcmd/TestJcmdDumpPathToGCRoots.java failed with error message: "Execution failed: `main' threw exception: java.lang.Exception: Expected chains but found none " #2766

We should revisit the ProblemList files and verify which jdk_jfr testcases are problematic and need to remain excluded, and see what upstream changes need to occur to stabilize the tests on various platforms to get them reenabled/fixed (and attempt to reduce the number of jfr testcases that need to be disabled).

Amalgamate and close the above list of issues, ideally by resolving the core issue, but at the very least by updating ProblemList files to refer to the correct upstream issue, or this one.

[sophia-guo]

Some failures might be obsolete. I'm going through each version to upload latest status.

For jdk21, the most problematic testcases fail on windows32 and arm32, both of which are not supported in jdk21 in adoptium #4983. Should we just ignore those failures https://github.com/adoptium/aqa-tests/blob/master/openjdk/excludes/ProblemList_openjdk21.txt#L421-L446 @smlambert

rerun jdk/jfr/jcmd/TestJcmdDump.java jdk/jfr/event/gc/detailed/TestZUncommitEvent.java on linux-x64, max-x64,windows-x64 - all passed
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8682/console
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8685/console
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8686/console

on all other platforms - all passed.
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8737/ ... https://ci.adoptium.net/view/Test_grinder/job/Grinder/8742/
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8745/
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8747/

will update the problemlist

[sophia-guo]

jdk17:
excluded on window32
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8688/
jdk/jfr/event/gc/collection/TestG1ParallelPhases.java.TestG1ParallelPhases failed

arm
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8751/ - all passed

[smlambert]

re: #4352 (comment) - yes, consider this list a snapshot in time. Feel free to close those related to jdk21 for 32-bit platforms.

We do however want to investigate stuff the other failures, noting that exclusion is typically a "temporary state", with the goal to solve issues and have tests re-included and running.

[roberttoyonaga]

Hi, I'll try to help out with this!

[sophia-guo]

@roberttoyonaga the list in description #4352 (comment) might be obsolete. You can also check the ProblemList_openjdk*. Problematic tests are excluded in the jdk_jfr sections in those problem list files.

[sophia-guo]

Rerun failed tests with jdk11 on windows32, ppc64le, linux-arm, all passed

https://ci.adoptium.net/view/Test_grinder/job/Grinder/8753/
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8754/
https://ci.adoptium.net/view/Test_grinder/job/Grinder/8755/

[sophia-guo]

jdk8: jdk/jfr/ on arm https://ci.adoptium.net/view/Test_grinder/job/Grinder/8756/

[sophia-guo]

@roberttoyonaga after the rerun it seems that most jfr test issues are with jdk8. See my recent run with jdk8 on arm32. https://ci.adoptium.net/view/Test_grinder/job/Grinder/8756/

#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGBUS (0x7) at pc=0xf62dd0fc, pid=2812, tid=0xd33e7460
#
# JRE version: OpenJDK Runtime Environment (8.0_402-b06) (build 1.8.0_402-b06)
# Java VM: OpenJDK Client VM (25.402-b06 mixed mode linux-aarch32 )
# Problematic frame:
# V  [libjvm.so+0x33b0fc]  write_checkpoint_header(unsigned char*, long long, long long, bool, unsigned int)+0xe8
#
# Core dump written. Default location: /home/jenkins/workspace/Grinder/aqa-tests/TKG/output_17073383472530/jdk_custom_0/work/scratch/2/core or core.2812
#
# An error report file with more information is saved as:
# /home/jenkins/workspace/Grinder/aqa-tests/TKG/output_17073383472530/jdk_custom_0/work/scratch/2/hs_err_pid2812.log
#
# If you would like to submit a bug report, please visit:
#   https://github.com/adoptium/adoptium-support/issues
#

[smlambert]

jdk/jfr/jcmd/TestJcmdDump.java - https://github.com/adoptium/aqa-tests/issues?q=is%3Aissue+TestJcmdDump - continues to fail across multiple platforms/versions intermittently with OOM during release triage. We should consider exclusion as it regularly occurs across a variety of our nodes.