-
Notifications
You must be signed in to change notification settings - Fork 278
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
v1.34.0 JVM crash #7144
Comments
Hi @zBart did you change anything else when you did the upgrade? Could you open a support ticket so I can get access to the hs_err file please? |
We have identified the cause of this bug and are working on releasing a fix |
I have similar symptoms. Is it the same reason?? I recently upgraded jdk from 17 to 23. jdk used: openjdk:23-jdk-slim dd-java-agent information: /datadog/dd-java-agent.jar 'https://dtdg.co/latest-java-tracer' Error message:
|
@foameraserblue there's nothing to suggest this was the same issue nor that it is related to the profiler. If you have the backtrace we would be able to determine if it's related or not. Since you're running an early access JDK23 build, it could always be a JVM problem. |
Hey Richard. Great! Do you still need me to send the hs_err file? |
We had a similar issue recently that affected some of our services running Java 21 . The same services had been running for months on Java 21, and years on Java 17, suddenly started crashing repeatedly during startup. We have a staging version of these services which were also experiencing the crashing but less reliably so.
We could not pinpoint a cause. We tried downgrading to Java 17 had the same issue. Had no sense of the cause, and due to our services running in AWS Fargate we were not able to access the error logs files Luckily we found these past issues which suggested possibly particularly this comment from @richardstartin :
Our build pulls the latest We then noticed that services which did not have profiling enabled were not crashing. Also, our staging services receive only 1% of the traffic of our prod services, and the Setting Going forward our workaround is to not using the profiling feature.
This is great to hear, would appreciate any updates to give us confidence on re-enabling the feature |
Hi @credpath-seek, this is unexpected. We test the profiler in lots of environments, and it runs in many other environments without these issues, so there is probably an incompatibility with something in your environment which we'd like to get to the bottom of to avoid this happening again. If we can get the backtrace and the siginfo sections from the hs_err files we would be able to pinpoint the cause. Just based on the error message, the cause is different to the crash reported (which is within AsyncGetCallTrace). If you do have this information available, please either reply here or open a support ticket so we can get the underlying issue fixed (and tested for). You could always re-enable profiling but set |
@zBart hs_err files are welcome. They help us identify how we can further protect the usage of AsyncGetCallTrace. Reproducers are even better, though I can see how that is hard to come by. Just to give more context on the profiling direction and what @richardstartin mentioned. We continue maintaining two flavours of profiling.
So by switching to the JFR events you will have a stable profiling experience. If you feel you are missing some features, we would be happy to get this feedback. |
Done, ticket ID is: #1749443 Any idea for an ETA for a fix? |
Hi @zBart - the fix is in DataDog/java-profiler#107 which will be coming to dd-trace-java in the upcoming release (beginning of July 2024) |
Is this fixed by #7229 ? |
meant asking if #7144 is fixed by #7229 Running latest Temurin 11 on g1 linux-aarch64
|
@oriy This is a different failure mode, unfortunately :( |
Got the same errors on the latest AWS Graviton CPU (
And there are no errors with disabled DD profiler by setting |
@ivan-sukhomlyn Hi, thanks for the report. This is a different failure mode - Would you be able to open a support ticket and submit the full hs_err.log file there so we can properly analyze the full crash? |
Sure @jbachorik, submitted ticket 1760710 |
Hi @jbachorik This is an answer from the support team regarding the requests mentioned in this issue.
|
We recently upgraded
dd-trace-java
from1.31.2
to1.34.0
and have gotten a crash. We've only seen this happens once so far.Not completely sure if this is a Datadog issue or a JVM issue.
Here is the log:
The text was updated successfully, but these errors were encountered: