RFR: 8308187: jdi/EventSet/resume/resume008 failed with "EventHandler> Unexpected event: ThreadStartEvent in thread resume008-thread0"

Chris Plummer cjplummer at openjdk.org
Thu May 18 22:12:57 UTC 2023


On Thu, 18 May 2023 15:49:22 GMT, Chris Plummer <cjplummer at openjdk.org> wrote:

> resume008 test is failing because of an unexpected Cleaner-1 thread being created:
> 
> ` # ERROR: debugger> ERROR: ThreadStartEvent is not for expected thread: Cleaner-1 `
> 
> Threads like Cleaner-1 are suppose to be filtered out so the test doesn't see them. Cleaner-1 had been missed by the filter since it never turned up in testing before. For some reason it started turning up 3 days ago. I'm not sure why, but it doesn't really matter since it should be filtered.
> 
> Testing
> 
> - Ran test 200 times. Before fix failed 13 times in 200 runs.
> - Ran  all the tier5 svc tests, which is where this issue has been turning up.

> The code that creates the "Cleaner-N" threads is from 202.02.04 so a very
long time ago. This failure first showed up in jdk-21+23-1950-tier5 and I can't
find anything remotely obvious that could have caused these failures to start
happening.

There are two different types of cleaner threads that can be created: `Cleaner-<n>` and `Common-Cleaner`. They have both been around for a long time. We'd seen the `Common-Cleaner` thread creation popup before during the debugger tests, but not `Cleaner-<n>.` Something changed, probably in the libraries, that is now triggering the creation of `Cleaner-<n>` when this test is run where previously it did not. But I don't think it is worth tracking down. The thread creation itself is not a bug. The bug is that it never made it into the filter list.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/14048#issuecomment-1553722101


More information about the serviceability-dev mailing list