Hi! It is a bug if the sleeping connection is in the auto-commit mode. But we need more information of the problem. If you encounter it, please send the exact sequence of SQL commands which leads to the problem.
You may also test SET AUTOCOMMIT=1 explicitly in your program. Note that LOCK TABLES switches auto-commit off until the table locks have been released. Regards, Heikki ........... Walt / Heikki Yes, I agree - with autocommit on, you do not need to supply COMMIT, and yes, that's the same as other RDBMSs. What I'm pointing out though is that I thought Heikki was suggesting that we need to supply the BEGIN/COMMIT for a single command even if autocommit was on, in order to not get the error "Lock wait timeout exceeded; Try restarting transaction". If that's the case, that would indicate that Mysql behaviour would be different to what I would expect. With auto-commit ON, and no explicit "LOCK TABLES' mentioned, I would not expect to get any error about a lock wait timeout unless the same ROW were being selected as were being updated, inserted or deleted. With a sleeping connection being the second potential row-locker, I would not expect to see this error ever. What I'm saying is that this looks like a bug. Sam mentioned that he only had one other connection when he got this error, and that this connection was a sleeping, persistent connection. The behaviour I saw was the same. A sleeping connection was supposedly locking a row needed (for a delete in my case). I don't think a sleeping connection (ie. one not actually running a query) should ever lock out a running query. I hope that clears up the confusion. John --------------------------------------------------------------------- Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php