Experiencing an issue with ScheduledExecutorService alongside VT

[Alan Bateman]

On 21/07/2024 14:11, Yuval Lombard wrote:
> Hi Alan,
>
> Attaching the thread dump when the issue occurs (as a txt file)
>
> It might be useful
>
Thanks. From a quick look, ForkJoinPool-1-worker-4 has an interesting 
stack trace. It tells us that a virtual thread has unmounted while 
trying to enter a monitor and the carrier is scheduling it to 
unblock+retry after a delay (this is detail on what the first thread to 
contend has to do in the object monitor implementation). It may be that 
the thread dump was taken at just the right thing to capture this 
contented on scheduling the wakeup but it may be something else. There 
is at least one issue in this area fixed in the more recent EA build so 
it would be good to try that. In the addition to the thread dumps, can 
you add `jcmd <pid> Thread.vthread_summary` to the list of things to 
capture? Asking because that will print a summary of timer queues which 
might be relevant to the issue.

-Alan