-
Notifications
You must be signed in to change notification settings - Fork 728
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
OpenJDK Regression suite "jdk_util" throws Segmentation Error before running tests on MacOS #5399
Comments
Crash in the GC iterating ClassHeapIterator.nextClassEv
fyi @dmitripivkine |
There is no results stored in these jobs. Would somebody help to get system core for this failure please? |
It fails different places, not just class iterating:
|
Notes from chatting with @dmitripivkine |
I presumed a connection based on their shared moment of failure. It happens during jdk_util test startup, before it has declared its running any tests. |
I was able to reproduce crash once however system core was not stored in results as well as was not discovered on machine directly in preserved workspace. So it is useless. It is new place of crash
|
Also I noticed another (much more common) failure:
|
Captured a core and got following stacktrace (similar with 3rd link in this issue):
Note: this crash occurs quite frequently and is easy to reproduce (in variable places though). However the core files just disappear when current test (not grinder) finishes. I have to use fyi @smlambert @llxia |
We do not yet capture core files from openjdk tests (dependent on this WIP PR adoptium/aqa-tests#1007), but thanks for letting me know of this failure that is repeatable, as I can use it to test my changes in that PR before merging. |
@smlambert good to hear that adoptium/aqa-tests#1007 is going to archive the diagnostic files. |
I looked to one of crashes. It crashed in Global GC an attempt to scan slot of String Table in There are eight hash tables in this String Table array. It crashed an attempt to scan very first slot for table
because of bad pointer (or memory corrupted content) for
This hash table is not converted to AVL tree so an address at Please contact me if you need more information or preserved system core(s) |
Was on other items instead. |
It doesn't look like we'll have a solution ready for this in the 0.15 release. |
Re: #5399 (comment) Hey @dmitripivkine, looks like a user in the OpenJ9 slack is seeing a problem that looks like this. Especially this bit:
https://openj9.slack.com/archives/C8312LCV9/p1565452794048200 |
Could this be related to this issue's original cause, or should I open a new issue to cover it? |
FYI, Dmitri is away for the week. |
@adamfarley Would you try to reproduce the problem with latest JVM? DDR is fixed now, so we can investigate system core |
Looking to information dumped before. Looks like most likely this is the same problem (the same pattern) as #5988 |
This should now be fixed, if there are tests excluded they can be unexcluded. |
To my knowledge, no tests were excluded as part of this bug. No tests will be reintroduced. |
Failure link
https://ci.adoptopenjdk.net/job/openjdk11_j9_openjdktest_x86-64_macos_macosXL/46/console
https://ci.adoptopenjdk.net/job/openjdk12_j9_openjdktest_x86-64_macos_macosXL/25/console
https://ci.adoptopenjdk.net/job/openjdk11_j9_openjdktest_x86-64_macos/213/console
Optional info
Last pass is
First failure is
8cbe269...127ba34
eclipse-openj9/openj9-omr@9f86693...5ceecf1
ibmruntimes/openj9-openjdk-jdk11@5650a71...ff6f49a
Failure output
The text was updated successfully, but these errors were encountered: