-
Notifications
You must be signed in to change notification settings - Fork 737
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
JDK 19 MiniMix_5m_1 hang in Thread.getStateImpl #16036
Comments
@ChengJin01 Please take a look at this |
As for the J9VMThread
|
There are a couple of javacore files, taken a minute apart, and both showed the same stack for the "load-9" thread. |
Assuming there was something happening for "load-9" thread according to the javacores,
it should be stuck at the native method
where the native code of openj9/runtime/jcl/common/thread.cpp Line 87 in 96fe9b8
and openj9/runtime/vm/VMAccess.cpp Line 1296 in e674448
|
Launched another Grinder (x200) at https://openj9-jenkins.osuosl.org/job/Grinder/1327/ with the latest passing build at https://openj9-jenkins.osuosl.org/job/Pipeline-Build-Test-JDK19/167/ to see how it goes. |
@ChengJin01 Do you know if this fails on JDK17? |
This job never fails on JDK17 in recent histories at https://openj9-jenkins.osuosl.org/view/Test_System_Nightly/job/Test_openjdk17_j9_extended.system_s390x_linux_Nightly/, but we can launch a Grinder with the latest build to double-check. |
Launched a Grinder (x200) at https://openj9-jenkins.osuosl.org/job/Grinder/1329/ with the latest build at https://openj9-jenkins.osuosl.org/view/Test_System_Nightly/job/Test_openjdk17_j9_extended.system_s390x_linux_Nightly/328/. |
The previously completed Grinder at https://openj9-jenkins.osuosl.org/job/Grinder/1327/ ended without any issue (killed by after 98 runs due to timeout) at https://openj9-jenkins.osuosl.org/job/Grinder_testList_0/161/consoleText
So it seems the hang was only triggered by the changes between https://openj9-jenkins.osuosl.org/job/Pipeline-Build-Test-JDK19/168/
and https://openj9-jenkins.osuosl.org/job/Pipeline-Build-Test-JDK19/167/
|
If I understand the above comments, this rules out #15227 as the cause (committed in June). It would be helpful to see the |
There was no hang detected in the Grinder at https://openj9-jenkins.osuosl.org/job/Grinder/1329/ for JDK17. So the problem is most likely related to Project Loom in JDK19 for the merged changes in OpenJ9 between c19e426 and 710fd95 given there is nothing suspicious in OMR changes. FYI: @fengxue-IS |
Nothing immediately suspicious - does the test fail with |
Just launched a couple of Grinders (x100) with the latest build on other platforms plus the failing build with |
So far we've only ever seen this hang once, in jdk19 nightly builds. |
All grinders at #16036 (comment) passed without any issue: So it seems the hang only occurred on Linux/s390x with JIT enabled and was captured only once in Grinder(x50) at https://openj9-jenkins.osuosl.org/job/Grinder/1321/consoleText. |
If that's the case, getting the information I requested earlier (thread flags, java stacks and native stacks) will help us figure out what's going on. |
#define J9_PUBLIC_FLAGS_HALT_THREAD_EXCLUSIVE 0x1 But the java stacktrace of thread
So I will revert the changes to see how it goes in Grinder. |
So far the problem can only be produced on the failing build at https://openj9-jenkins.osuosl.org/job/Grinder/1408/consoleText as I tried both with the latest build and with the reverted change at https://github.com/ChengJin01/openj9/tree/revert_jit_zlinux_thrd_pin_15752 in which case the hang never occurred. I think the JIT team still need to take a look to see what happened in there. FYI: @r30shah |
The impression I get from @pshipton is that this happened exactly once and has never been reproduced, so I'm not sure there's any reason to blame any change. Was the pinning support updated in the next build? |
As I mentioned above, the hang is only reproduced on the failing build but never on the build right before the failing build or on the latest build.
I didn't see any update on these two files in #15752 or anything related to the pinning support after it was merged. |
Just talked to GAC, we close the issue given there is no way to reproduce it in the latest build and no further evidence to prove the hang was caused by any change unless the problem shows up again. |
https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_extended.system_s390x_linux_Nightly_testList_1/19
MiniMix_5m_1
https://openj9-artifactory.osuosl.org/artifactory/ci-openj9/Test/Test_openjdk19_j9_extended.system_s390x_linux_Nightly_testList_1/19/system_test_output.tar.gz
The text was updated successfully, but these errors were encountered: