-
Notifications
You must be signed in to change notification settings - Fork 3.6k
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
[Bug] JVM crash in unit tests: PhaseIdealLoop::build_loop_late_post_work #19307
Comments
The issue had no activity for 30 days, mark with Stale label. |
I found this problem in the test environment. |
The issue had no activity for 30 days, mark with Stale label. |
Closed according to #19307 (comment). If it comes back, feel free to open a new issue. |
Hi, Search before asking
VersionMinimal reproduce step
What did you expect to see?Messages produced with success. Brokers do not restart. What did you see instead?After around 30 000 messages, Brokers keep restarting with the following error:
Important logs:
Anything else?Seems like exactly the bug https://bugs.openjdk.org/browse/JDK-8285835?attachmentViewMode=list which is fixed in JDK
Seems like releasing a new Docker image with newer JDK should fix the problem (I've noticed that Ubuntu version was upgarded). Probably building the image right now should download the latest Would that be possible to release bugfix image with never JDK? Thank you. |
it happened again in pulsar-metadata tests.
https://gist.github.com/lhotari/e191a541ca4558b624736b27620f48cd contains the full error report file. |
this issue still has label stale, see #21458 |
More recent crashes with similar stack trace: https://gist.github.com/lhotari/53b72683ad4f339dfbcfd8b9b97062b9 |
It looks like the issue will be fixed in 17.0.10 with https://bugs.openjdk.org/browse/JDK-8315861 which is the backport of https://bugs.openjdk.org/browse/JDK-8314024 . |
The fix https://bugs.openjdk.org/browse/JDK-8314024 is already included in 21.0.1 which has already been released. One possibility would be to run Pulsar on Java 21.0.1 to get the fix earlier. An ea build of 17.0.10 is another option. That should be available soon. The third option is to wait until 17.0.10 is released in the last half of January. |
There's another JVM crash where it happens with a different type of stack trace, in RocksDB:
reported as #19250 |
Search before asking
Version
9e698b6
Minimal reproduce step
none, this happens with pulsar-metadata tests. Perhaps related to #19250
What did you expect to see?
No JVM crashes
What did you see instead?
The JVM crashed and caused the build to fail
The hs err file contains:
Anything else?
Full hs err file: https://gist.github.com/lhotari/4948f347a8f70f6a1051212139fcf551
another one: https://gist.github.com/lhotari/f3892d3577e285edc469141577215261
Failed build: https://github.com/lhotari/pulsar/actions/runs/3985274820/jobs/6832729920#step:11:1655
Are you willing to submit a PR?
The text was updated successfully, but these errors were encountered: