Open JasonFengJ9 opened 2 years ago
# DETECTED_JAVA_VERSION=openjdk version "11.0.24-internal" 2024-07-16
# OpenJDK Runtime Environment (build 11.0.24-internal+0-adhoc..BuildJDK11s390xlinuxjitPersonal)
# Eclipse OpenJ9 VM (build master-abc6fa4ae38, JRE 11 Linux s390x-64-Bit Compressed References 20240711_1485 (JIT enabled, AOT enabled)
# OpenJ9 - abc6fa4ae38
# OMR - 5e425ebfc86
# JCL - fa2af551c3d based on jdk-11.0.24+7)
# DETECTED_RELEASE_INFO=IMPLEMENTOR="Eclipse OpenJ9"
# JAVA_RUNTIME_VERSION="11.0.24-internal+0-adhoc..BuildJDK11s390xlinuxjitPersonal"
# JAVA_VERSION="11.0.24"
# JAVA_VERSION_DATE="2024-07-16"
# LIBC="gnu"
# OS_ARCH="s390x"
# OS_NAME="Linux"
# SOURCE="OpenJDK:fa2af551c3d OpenJ9:abc6fa4ae38 OMR:5e425ebfc86"
#
# Hostname: rhel9s390x-svl-rt10-1.fyre.ibm.com
not ok 165 - testSoftMxNotDisclaimMemory_zlinux_64_3
---
output:
|
===============================================
===============================================
testSoftMxNotDisclaimMemory_zlinux_64_3 Start Time: Thu Jul 11 02:11:17 2024 Epoch Time (ms): 1720689077502
variation: Mode501
JVM_OPTIONS: -XX:+UseJITServer -Xjit -Xgcpolicy:balanced -Xnocompressedrefs
TEST SETUP:
JVMSHRC005I No shared class caches available
JVMSHRC005I No shared class caches available
cache cleanup done
TESTING:
[IncludeExcludeTestAnnotationTransformer] [INFO] EXCLUDE_FILE environment variable: /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/../TestConfig/resources/excludes/latest_exclude_11.txt
[IncludeExcludeTestAnnotationTransformer] [INFO] Processing exclude file: /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/../TestConfig/resources/excludes/latest_exclude_11.txt
...
... TestNG 6.14.2 by Cédric Beust (cedric@beust.com)
...
[SoftmxAdvanceTest] [WARN] Generate Java core dump after forcing GC.
JVMDUMP034I User requested Java dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/javacore.20240711.021622.3274543.0001.txt' through com.ibm.jvm.Dump.JavaDump
JVMDUMP010I Java dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/javacore.20240711.021622.3274543.0001.txt
JVMDUMP034I User requested Heap dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/heapdump.20240711.021622.3274543.0002.phd' through com.ibm.jvm.Dump.HeapDump
JVMDUMP010I Heap dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/heapdump.20240711.021622.3274543.0002.phd
JVMDUMP034I User requested System dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/core.20240711.021623.3274543.0003.dmp' through com.ibm.jvm.Dump.SystemDump
JVMDUMP010I System dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/core.20240711.021623.3274543.0003.dmp
[SoftmxAdvanceTest] [ERROR] FAIL: Heap can't shrink to the reset max heap size within 5 minutes!
[SoftmxAdvanceTest] [WARN] Generate Java core dump after forcing GC.
JVMDUMP034I User requested Java dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/javacore.20240711.022125.3274543.0004.txt' through com.ibm.jvm.Dump.JavaDump
JVMDUMP010I Java dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/javacore.20240711.022125.3274543.0004.txt
JVMDUMP034I User requested Heap dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/heapdump.20240711.022125.3274543.0005.phd' through com.ibm.jvm.Dump.HeapDump
JVMDUMP010I Heap dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/heapdump.20240711.022125.3274543.0005.phd
JVMDUMP034I User requested System dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/core.20240711.022125.3274543.0006.dmp' through com.ibm.jvm.Dump.SystemDump
JVMDUMP010I System dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/core.20240711.022125.3274543.0006.dmp
[SoftmxAdvanceTest] [ERROR] FAIL: Heap can't shrink to the reset max heap size within 5 minutes!
[SoftmxAdvanceTest] [ERROR] FAIL: Memory shouldn't release back to OS! Post memory size 1695461376 is greater than previous memory size 1607172096
[SoftmxAdvanceTest] [ERROR] FAIL: Memory shouldn't release back to OS! Post memory size 1734246400 is greater than previous memory size 1694412800
[SoftmxAdvanceTest] [WARN] Generate Java core dump after forcing GC.
JVMDUMP034I User requested Java dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/javacore.20240711.022642.3274543.0007.txt' through com.ibm.jvm.Dump.JavaDump
JVMDUMP010I Java dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/javacore.20240711.022642.3274543.0007.txt
JVMDUMP034I User requested Heap dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/heapdump.20240711.022642.3274543.0008.phd' through com.ibm.jvm.Dump.HeapDump
JVMDUMP010I Heap dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/heapdump.20240711.022642.3274543.0008.phd
JVMDUMP034I User requested System dump using '/home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/core.20240711.022642.3274543.0009.dmp' through com.ibm.jvm.Dump.SystemDump
JVMDUMP010I System dump written to /home/jenkins/workspace/Test_openjdk11_j9_extended.functional_s390x_linux_jit_Personal_testList_1/aqa-tests/TKG/output_17206828973948/testSoftMxNotDisclaimMemory_zlinux_64_3/core.20240711.022642.3274543.0009.dmp
[SoftmxAdvanceTest] [ERROR] FAIL: Heap can't shrink to the reset max heap size within 5 minutes!
FAILED: testDisclaimMemoryEffects(5)
java.lang.AssertionError: Test has failed after 5 attempts
This has started failing somewhat frequently in the JITServer nightly tests again, e.g. here: https://hyc-runtimes-jenkins.swg-devops.com/job/Test_openjdk8_j9_extended.functional_s390x_linux_jit_Personal_testList_1/1084/ on rhel8s390x-svl-rt5-1.fyre.ibm.com. Normally I'd think that maybe the use of JITServer is making the failure happen more frequently, but this test runs with -Xnocompressedrefs
, so the JITServer shouldn't have been in use during these failures.
This test was unreliable from the very beginning. Please note there are a few aspects of this test:
mmap()
Port Library calls madvise()
with "don't need" request. If memory pages have been reserved with shmat()
(pre-allocated shared Huge pages) Port Library does nothing, so there is no memory returned back to OS and test fails.madvise()
with "don't need" does not guarantee that returned memory is going to be available as "free" in OS also. It really depends of working environment. Ex. Hyperviser (if used) might decide to keep memory around instead of return to OS. OS itself might decide do not claim memory back to free etc.Another (completely orthogonal) point: please note that there are two Indexable objects headers types selected dynamically run-time now. Standard collectors (Gencon, Optthruput and Opeavgpause) and Metronome use one of them (default) and Balanced uses another one (extended by dataAddr
field). It makes code generated for Gencon be incompatible for Balanced. I don't know can this fact be related to these failures.
And GC verbose log from your build shows requested heap contraction instantly:
<mem-info id="3711" free="30408704" total="444596224" percent="6">
<mem-info id="3718" free="30283464" total="444596224" percent="6">
<mem-info id="3725" free="266338304" total="268435456" percent="99"> <---
<mem-info id="3733" free="266292280" total="268435456" percent="99">
So, GC works as expected. There are many factors can make this test fail. I think non of them relates to JVM directly (but test running environment).
Failure link
From an internal build(
rhel7s390x-3-1
):Rerun in Grinder - Change TARGET to run only the failed test targets.
Optional info
Failure output (captured from console output)
Related to https://github.com/eclipse-openj9/openj9/issues/4475