xc-jianghan commented on issue #4814:
URL: https://github.com/apache/kyuubi/issues/4814#issuecomment-3864242522
Thank you very much for taking the time to reply to a student like me
despite your busy schedule. Your judgment makes a lot of sense. When I was
using AI to help me learn, it also suggested the same direction. But since
you’re the creator, you understand this architecture better than anyone.
Following the AI’s suggestions, I made several changes and recompiled
multiple times. The main changes were:
1.
removing the lazy initialization, and
2.
creating the AtlasClient earlier.
However, neither approach was able to solve the issue.
My current environment is Hadoop 3.4.0 + Spark 3.5.3 + Atlas 2.4.0. Could
you give me some guidance? I’m willing to pay for consulting, and I would be
very grateful for any advice you can share.
Cheng ***@***.***> 在 2026年2月7日 周六 18:59 写道：
pan3793 left a comment (apache/kyuubi#4814)
Oh, interesting, the reason should be: AtlasClient is initialized lazily,
and it will add a shutdown hook during initialization.
but if we catch and swallow the exception on registerCleanupShutdownHook,
AtlasClient may not be able to close properly and results to leak (e.g.,
non-daemon threads might blocks driver JVM from exiting), so we may need to
find another way to init the AtlasClient earlier
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you commented.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
