Ok, I suspected it might be an exception, but I was surprised the abort() occurred on the background thread. I thought it was the main thread that typically performed the final abort() due to an exception that occurred on any thread, but now its clear that the abort() is actually called on the exception thread.

It sounds like you are experiencing the same bug we are experiencing. This bug still exists as of latest Sequoia release. Disturbing.

Has anyone started a new thread about this? What is the status of this apple bug? We have written 2 bug reports to apple with no response yet. Please advise.