After a question I asked, it appears the malfunction in our application comes from a bug in the JVM we are using. It is a bug causing a segmentation fault in the Java2d disposal thread, and the Java bugtracker indicates it could be caused by the deletion of a file while the JVM is accessing it. The bug is identified, but I have a hard time to confirm it is actually solved, and in which version of Java. It is simply marked as "closed" Oracle's bugtracker, with a reference to another issue, which isn't available online anymore (probably a consequence of the Sun-Oracle merger and the disruption it caused in official Java-related websites). Could anyone help?
Asked
Active
Viewed 119 times
1
-
1You might also want to paraphrase your earlier question, in perhaps 2-3 sentences just for completeness +1. – Tim Biegeleisen Dec 20 '18 at 13:56
-
Make an explicit test that this bug is not present in the JVM your application is running on. – Thorbjørn Ravn Andersen Dec 20 '18 at 13:58
-
@ThorbjørnRavnAndersen The bug IS present in the JVM I'm using. And the bug is not easy to reproduce. – Alexis Dufrenoy Dec 20 '18 at 14:00
-
Why the downvote? – Alexis Dufrenoy Dec 20 '18 at 14:00
-
Nice, in the link [bug is identified](https://bugs.java.com/bugdatabase/view_bug.do?bug_id=7103530) in your question they mention "*EVALUATION This is certainly a duplicate of 6953445.*" however when trying to access this bug (to see comments/how it was resolved) it says "*This bug is not available.*" – Mark Dec 20 '18 at 14:02
-
@Mark This most likely mean that the duplicate is a private one which outsiders cannot access. – Thorbjørn Ravn Andersen Dec 20 '18 at 14:05
-
@Mark: That's what I meant by "a reference to another issue, which isn't available online anymore". Maybe it never was... – Alexis Dufrenoy Dec 20 '18 at 14:31
-
Consider editing your question and including your code which deletes temp files. Are you being careful to only delete temp files which your program explicitly created? – VGR Dec 20 '18 at 15:31
-
I don't have code deleting temp files. Or at least not identified code. We are looking into other applications interfering or even manual operations, to no avail, yet. – Alexis Dufrenoy Dec 20 '18 at 15:53
1 Answers
3
This is not a bug in the JDK that has been fixed in a given version. The evaluation at the bottom says:
This is certainly a duplicate of 6953445.
User deleting temp files being actively used by the JDK.
This means that your application removes files that the JDK expects to be there, so my guess is the resolution is that this is not a bug in the JDK but your code.
If you really MUST know then Oracle support (most likely paid, unless you have a license from other products) should be able to give a definitive answer.
Community
- 1
- 1
Thorbjørn Ravn Andersen
- 73,784
- 33
- 194
- 347
-
So you mean it's normal the JVM is crashing because of a temp file deletion? It's not considered a bug? – Alexis Dufrenoy Dec 20 '18 at 14:27
-
What _I_ mean doesn't matter. What matters is that the engineer who looked at said bug report said that it isn't a bug and won't be changed. Fix your code. – Thorbjørn Ravn Andersen Dec 20 '18 at 14:30
-
I'm trying to solve my problem. Like many people noticed, the #7103530 issue indicates a reference to #6953445, so things are not as clear as you're stating. Heck, I'm not even sure I'm looking at the right bug and that's even related to what's happening in my application. You're not helping. So if you don't have any valuable information, feel free to stop posting! – Alexis Dufrenoy Dec 20 '18 at 14:39
-
Naturally I'm guessing. Duly noted that you don't consider it valuable information that this behavior is most likely unchanged in _any_ newer version of Java. To pursue this further I would suggest looking at open file handles while your JVM is running - that is available in the appropriate "/proc" subdirectory. – Thorbjørn Ravn Andersen Dec 20 '18 at 14:44
-
My app crashes something like once in 6 to 8 months. Sorry for not repeating ALL of my previous question here, but I'm not sure moderators would be very happy about it if I did. You say that behavior is "most likely" unchanged in newer versions, but how do you know? Do you have access to issue #6953445? – Alexis Dufrenoy Dec 20 '18 at 14:47
-