-
Notifications
You must be signed in to change notification settings - Fork 723
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
jdk20 OpenJDK java/util/concurrent/locks/Lock/OOMEInAQS OOM crash vmState=0x00040000 #16659
Comments
5x grinder https://openj9-jenkins.osuosl.org/job/Grinder/1984/ - passed 10 iterations, 5 each in modes 0, 1. |
https://openj9-jenkins.osuosl.org/job/Test_openjdk20_j9_sanity.openjdk_x86-64_linux_Nightly/5
|
@summary Check that AQS-based locks, conditions, and CountDownLatches do not fail when encountering OOME The test is new for jdk20, and the OOM is intentional. |
https://openj9-jenkins.osuosl.org/job/Test_openjdk20_j9_sanity.openjdk_s390x_linux_Nightly/26
|
JDK20 aarch64_linux(
|
FYI: I was wondering why it goes OOM without heap expansion,
|
The OOM is intentional [1].
|
@tajila can VM start the investigation on this one and see why the crash occurs. |
See the crash in #16659 (comment) with a stack trace. |
The segfault is caused because a thread being interrupted is being shutdown: Thread 0x000000000E1486D8 "Attach API wait loop" is already shutdown when "Attach API teardown" attempts to interrupt it
|
|
The wait loop thread (j9vmthread 0x000000000124C100) is in the vm->deadthreadlist |
Grinder with more method tracing on zlinux https://openj9-jenkins.osuosl.org/job/Grinder/2253/ |
No cores in previous grinder, rebuilding https://openj9-jenkins.osuosl.org/job/Grinder/2257/ |
Started some grinders with better options, the earlier ones didn't run the test. It is running the test standalone and things don't always fail that way if the JIT is involved. |
It failed in the first grinder, but I was stupid and didn't manually copy the core and other artifacts before stopping the run. |
Related to: eclipse-openj9/openj9#16659 Signed-off-by: tajila <[email protected]>
Related to: eclipse-openj9/openj9#16659 Signed-off-by: tajila <[email protected]>
Related to: eclipse-openj9/openj9#16659 Signed-off-by: tajila <[email protected]>
The JDK21 changes have been merged. I can't confirm that it fixes the issues given that it is quite intermittent. |
I'll close it and then we can reopen if it's seen again. |
Related to: eclipse-openj9#16659 Signed-off-by: tajila <[email protected]>
https://openj9-jenkins.osuosl.org/job/Test_openjdk21_j9_sanity.openjdk_s390x_linux_Nightly_testList_1/133
|
The test forces OOME's so the following is expected:
I think the only issue is that it is running slow, likely due to excessive GC's, which causes the timeout. The timeout handler is also slow for the same reasons. This is an extreme test case, so I dont think it should block the release. |
@TobiAjila we can't have this test intermittently failing all the time. Should we exclude it or can it be fixed to be less aggressive? |
The test is designed to stress OOM paths. I think now that we've fixed the segfault we can disable this test. Im not sure how to modify it to be less agressive. |
Ill open a PR to exclude it |
I suppose this is perm excluded. Changing the label and closing. |
https://openj9-jenkins.osuosl.org/job/Test_openjdk20_j9_sanity.openjdk_s390x_linux_OpenJDK20/1
jdk_util.0
java/util/concurrent/locks/Lock/OOMEInAQS.java
The text was updated successfully, but these errors were encountered: