Re: RFR(XS): 8209029 ProblemList tests that fail due to 'Error attaching to process: Can't create thread_db agent!' in jdk-11+25 testing

[Daniel D. Daugherty]

On 8/6/18 8:39 PM, David Holmes wrote:

Hi Dan,

Reviewed.

Thanks!

Dan

As discussed on IM the id#1 test is for ZGC and we don't support ZGC on Solaris so no point excluding it. :)

Thanks,
David

PS. Yes I will file jtreg RFE.

On 7/08/2018 10:14 AM, Daniel D. Daugherty wrote:

On 8/6/18 7:46 PM, David Holmes wrote:

Hi Dan,

On 7/08/2018 6:53 AM, Daniel D. Daugherty wrote:

Greetings,

I'm in the process of reducing the noise in the JDK11 and JDK12 CIs so
I need a single (R)eviewer for the following fix:

   JDK-8209029 ProblemList tests that fail due to 'Error attaching to

               process: Can't create thread_db agent!' in jdk-11+25 testing

   https://bugs.openjdk.java.net/browse/JDK-8209029

Yes, this is the third sub-task for ProblemListing test failures due
to this issue.

Here's the diff:

$ hg diff
diff -r f3e15dcdc333 test/hotspot/jtreg/ProblemList.txt

--- a/test/hotspot/jtreg/ProblemList.txt    Mon Aug 06 00:46:40 2018 -0700 +++ b/test/hotspot/jtreg/ProblemList.txt    Mon Aug 06 16:28:16 2018 -0400

@@ -85,8 +85,10 @@

  serviceability/sa/ClhsdbAttach.java 8193639 solaris
  serviceability/sa/ClhsdbCDSCore.java 8207832 linux-x64
+serviceability/sa/ClhsdbCDSJstackPrintAll.java 8193639 solaris
  serviceability/sa/ClhsdbField.java 8193639 solaris
  serviceability/sa/ClhsdbFindPC.java 8193639 solaris
+serviceability/sa/ClhsdbFlags.java 8193639 solaris
  serviceability/sa/ClhsdbInspect.java 8193639 solaris
  serviceability/sa/ClhsdbJdis.java 8193639 solaris
  serviceability/sa/ClhsdbJhisto.java 8193639 solaris
@@ -113,9 +115,13 @@
  serviceability/sa/TestDefaultMethods.java 8193639 solaris
  serviceability/sa/TestG1HeapRegion.java 8193639 solaris
  serviceability/sa/TestHeapDumpForInvokeDynamic.java 8193639 solaris
+serviceability/sa/TestHeapDumpForLargeArray.java 8193639 solaris

+serviceability/sa/TestInstanceKlassSizeForInterface.java 8193639 solaris

+serviceability/sa/TestIntConstant.java 8193639 solaris
+serviceability/sa/TestJhsdbJstackLock.java 8193639 solaris

  serviceability/sa/TestRevPtrsForInvokeDynamic.java 8191270 generic-all

  serviceability/sa/TestType.java 8193639 solaris
-serviceability/sa/TestUniverse.java 8193639 solaris
+serviceability/sa/TestUniverse.java#id0 8193639 solaris

  #############################################################################

The entry for serviceability/sa/TestUniverse.java did not prevent
that test from running so I'm adding the '#id0' part of the name
that shows up in JTREG output about the failure. The source file
is called 'serviceability/sa/TestUniverse.java' which is why that
name was used in the previous version of the entry.

I chatted to Jon Gibbons about this. We should file a RFE to allow all the test cases in a "multi-test" to be excluded together.

I presume that you or Jon are going to file that RFE...

But, given the difference between #id0 and #id1 is whether ZGC is used or not, I would not expect the ZGC case to pass more often than the non-ZGC case. So it seems to me you will want to exclude both sub-tests here.

Our current plan with CI noise reduction is that we only ProblemList a
test when we have observed a failure. These groups of failures have
begun to stretch our patience since we are now on our 3rd ProblemList
update for JDK-8193639...

Also the comments at the top of the file indicate that my previous
entries should have been 'solaris-all' instead of just 'solaris' so
I'll make that change before I push.

Right.

So may I count this as a review?

Dan

Thanks,
David

Thanks, in advance, for any questions, comments or suggestions.

Dan