Bug ID: JDK-8343893 Test jdk/jfr/event/runtime/TestNativeMemoryUsageEvents.java failed: heap should have grown and NMT should show that: expected 0

Type: Bug
Component: hotspot
Sub-Component: runtime
Affected Version: 24

Priority: P4
Status: New
Resolution: Unresolved

Submitted: 2024-11-10
Updated: 2024-11-11

Test: jdk/jfr/event/runtime/TestNativeMemoryUsageEvents.java

----------System.out:(13/869)----------
[0.040s][info][gc] Using The Z Garbage Collector
[1.109s][info][gc] GC(0) Major Collection (Warmup)
[1.217s][info][gc] GC(0) Major Collection (Warmup) 14M(11%)->8M(6%) 0.107s
[1.385s][info][gc] GC(1) Major Collection (Warmup)
[1.506s][info][gc] GC(1) Major Collection (Warmup) 26M(20%)->82M(64%) 0.121s
[1.510s][info][gc] GC(2) Major Collection (Warmup)
[1.610s][info][gc] GC(3) Minor Collection (Allocation Rate)
[1.871s][info][gc] GC(3) Minor Collection (Allocation Rate) 80M(62%)->82M(64%) 0.260s
[1.878s][info][gc] GC(4) Minor Collection (Allocation Rate)
[1.937s][info][gc] GC(4) Minor Collection (Allocation Rate) 82M(64%)->86M(67%) 0.060s
[1.949s][info][gc] GC(5) Minor Collection (Allocation Rate)
[2.050s][info][gc] GC(5) Minor Collection (Allocation Rate) 86M(67%)->82M(64%) 0.101s
[2.050s][info][gc] GC(2) Major Collection (Warmup) 82M(64%)->82M(64%) 0.540s
----------System.err:(14/1012)----------
java.lang.RuntimeException: heap should have grown and NMT should show that: expected 0 > 0
	at jdk.test.lib.Asserts.fail(Asserts.java:691)
	at jdk.test.lib.Asserts.assertGreaterThan(Asserts.java:400)
	at jdk.jfr.event.runtime.TestNativeMemoryUsageEvents.verifyHeapGrowth(TestNativeMemoryUsageEvents.java:155)
	at jdk.jfr.event.runtime.TestNativeMemoryUsageEvents.main(TestNativeMemoryUsageEvents.java:188)
	at java.base/jdk.internal.reflect.DirectMethodHandleAccessor.invoke(DirectMethodHandleAccessor.java:104)
	at java.base/java.lang.reflect.Method.invoke(Method.java:567)
	at com.sun.javatest.regtest.agent.MainWrapper$MainTask.run(MainWrapper.java:138)
	at java.base/java.lang.Thread.run(Thread.java:1576)

Failed in tier8 on three platforms: linux-x64, linux-aarch64, windows-x64

The usage event looks like: jdk.NativeMemoryUsage { startTime = 19:15:07.214 (2024-11-08) type = "Java Heap" reserved = 16.1 GB committed = 0.0 MB } while NTM memory report has: Native Memory Tracking: Total: reserved=36038064519, committed=1758534023 malloc: 121268615 #153665, peak=120511943 #144555 mmap: reserved=35916795904, committed=1637265408 - Java Heap (reserved=17246978048, committed=1493172224) (mmap: reserved=17246978048, committed=1493172224, at peak) ..... The native memory report use: void MemBaseline::baseline_summary() { MallocMemorySummary::snapshot(&_malloc_memory_snapshot); VirtualMemorySummary::snapshot(&_virtual_memory_snapshot); { MemoryFileTracker::Instance::Locker lock; MemoryFileTracker::Instance::summary_snapshot(&_virtual_memory_snapshot); } _metaspace_stats = MetaspaceUtils::get_combined_statistics(); } and JFR inly virtual memory void NMTUsage::update_vm_usage() { const VirtualMemorySnapshot* vms = VirtualMemorySummary::as_snapshot(); // Reset total to allow recalculation. _vm_total.committed = 0; _vm_total.reserved = 0; for (int i = 0; i < mt_number_of_tags; i++) { MemTag mem_tag = NMTUtil::index_to_tag(i); const VirtualMemory* vm = vms->by_type(mem_tag); _vm_by_type[i].reserved = vm->reserved(); _vm_by_type[i].committed = vm->committed(); _vm_total.reserved += vm->reserved(); _vm_total.committed += vm->committed(); } } so it is needed to etiher update ZGC JFR reporting to track memory commited by MemoryFileTracker or to updated test to exepct that cmmited memory is zero.
11-11-2024
It is not lilliput problem. The test doesn't work with all GCs. It checks committed memory but for ZGC it doesn't add memory allocated by MemoryFileTracker.
11-11-2024
I will file a seperate bug, for tracking purposes, for the jdk/jfr/threading/TestDeepVirtualStackTrace.java failures.
11-11-2024
jdk/jfr/threading/TestDeepVirtualStackTrace.java also seems to fail with -XX:+UseCompactObjectHeaders and "java.lang.RuntimeException: No event found in virtual thread: expected false, was true".
11-11-2024
This happens with -XX:+UseCompactObjectHeaders. Roman has agreed to take a look at it.
11-11-2024
Moving to Runtime to investigate if there has been changes to NMT.
11-11-2024