Hello,

Some time ago, I reported a problem that sent the mysql service's CPU utilization to >99%. I have now been able to demonstrate the same problem on another machine by doing a direct copy of the data directory. Sadly, the "reproduceable case" is useless because: (1) it involves more records than it should; (2) so far, it is based on data that I can't release due to confidentiality constraints.

Still, being able to move it off of my production machine is a big help, because now I can more freely experiment, and it shows that it's not just the one box. Also, I can now say that Dolphin Smalltalk and MyODBC are not required: a batch file driving mysql is sufficient to clobber the server.

Any thoughts on how to distill the real database down to a simple example that illustrates the problem? So far, my guesses re the cause have not lead to failure. For example, I tried limiting the Access database to records that would be returned had my failed query worked. Then I imported that with MySQL-front (which is how I created the full database), but sadly the query works with the resulting database.

Bill

Wilhelm K. Schwab, Ph.D. bil...@anest4.anest.ufl.edu (352) 846-1285