TestJcmd testMatchDisplayNamePartially Expected string not found about openj9

[pshipton]

https://openj9-jenkins.osuosl.org/job/Test_openjdk17_j9_extended.functional_x86-64_windows_Nightly_testList_0/680
TestJcmd_0

23:34:00  FAILED: testMatchDisplayNamePartially
23:34:00  java.lang.AssertionError: Expected string not found: java.util.HashMap expected [true] but found [false]
23:34:00  	at org.testng.Assert.fail(Assert.java:96)
23:34:00  	at org.testng.Assert.failNotEquals(Assert.java:776)
23:34:00  	at org.testng.Assert.assertTrue(Assert.java:44)
23:34:00  	at org.openj9.test.attachAPI.TestJcmd.testDisplayNameHelper(TestJcmd.java:372)
23:34:00  	at org.openj9.test.attachAPI.TestJcmd.testMatchDisplayNamePartially(TestJcmd.java:383)

[pshipton]

@JasonFengJ9 I assume this is caused by the following, can you pls take a look
#19069
#19089

[JasonFengJ9]

JDK11 50x grinder - https://openj9-jenkins.osuosl.org/job/Grinder/3384/ - 4/50 failed
JDK17 50x grinder - https://openj9-jenkins.osuosl.org/job/Grinder/3385/ - 3/50 failed

[JasonFengJ9]

The test in question TestJcmd.java invokes TargetManager to create a target VM, then starts another process for Jcmd with corresponding arguments.

openj9/test/functional/Java8andUp/src/org/openj9/test/attachAPI/TestJcmd.java

Lines 357 to 367 in f2c434f

	TargetManager tgt = new TargetManager(TestConstants.TARGET_VM_CLASS, null, DISPLAYNAME_GC_CLASS_HISTOGRAM,
	Collections.singletonList("-Xmx10M"), Collections.emptyList());
	tgt.syncWithTarget();
	String targetId = tgt.targetId;
	assertNotNull(targetId, ERROR_TARGET_NOT_LAUNCH);
	List<String> args = new ArrayList<>();
	args.add(targetName);
	args.add(GC_CLASS_HISTOGRAM);
	args.add("all");
	List<String> jcmdOutput = runCommandAndLogOutput(args);

From the AttachAPI log files, the target VM was started successfully but exits promptly w/o responding an attaching request. The instrumental log message from the target VM didn't appear in the test console output, will add tracepoints to understand why it didn't wait for the attaching request.
This seems a test issue.

[JasonFengJ9]

The attaching failed due to a SocketException:

java.net.SocketException: Software caused connection abort: socket write error
        at java.base/java.net.SocketOutputStream.socketWrite0(Native Method)
        at java.base/java.net.SocketOutputStream.socketWrite(SocketOutputStream.java:110)
        at java.base/java.net.SocketOutputStream.write(SocketOutputStream.java:129)
        at java.base/openj9.internal.tools.attach.target.AttachmentConnection.streamSend(AttachmentConnection.java:105)
        at jdk.attach/com.ibm.tools.attach.attacher.OpenJ9VirtualMachine.executeDiagnosticCommand(OpenJ9VirtualMachine.java:318)
        at jdk.jcmd/openj9.tools.attach.diagnostics.attacher.AttacherDiagnosticsProvider.executeDiagnosticCommand(AttacherDiagnosticsProvider.java:55)
        at jdk.jcmd/openj9.tools.attach.diagnostics.tools.Util.runCommandAndPrintResult(Util.java:78)

This error in question is https://learn.microsoft.com/en-us/previous-versions/ms832256(v=msdn.10)?redirectedfrom=MSDN

Re-trying the test (3 times) passed consistently in 200x grinder.
Jcmd might be improved with re-tries in case this SocketException.
The failing test is a new test added recently via

• #19069

This behaviour is not a regression and will be addressed in 0.46.