Has anyone seen an issue with recent versions of Lucee 5.3 where joining all threads created still has some threads in a running state and the timeout has not be hit? We recently migrate our application from ACF to Lucee and in one of our background jobs, we spawn a bunch of threads to check some email mailboxes. What I'm seeing is that occassionally some of the threads are still in a "RUNNING" state after the join, even though the join is well short of the timeout. The join should be only timing out after about 15 minutes, but in these cases the join happens after about 30-60 seconds and some threads are still running. Someone else posted a message in the forums asking about join/running states just a few days ago: https://dev.lucee.org/t/threads-still-in-running-status-after-join/10644/3 Maybe the issue is within Java. I'm not seeing anything in the Lucee source code that would look like an obvious issue.

Don't know if it's related or not, but we have for some time been experiencing problems with threads spawned by using the

parallel

attribute of

.Each()

functions. Intermittently a thread will just hang there until we kill it using FusionReactor. Doesn't happen with all use of

parallel

, just sometimes with some uses.

We're not currently using that, but I appreciate the heads up. From what I can tell, the thread seems to have actually run to completion, but is just left in the "RUNNING" state. Either that, or it's finishing up just before we terminate it, but after the join finished waiting. Either way, the join should have continued until the state was "COMPLETED"

That sounds similar, i.e. completed but still running.