Victoria Reznichenko wrote: >Steve, >Wednesday, July 10, 2002, 11:12:48 AM, you wrote: > >SH> The bug can be reproduced as follows: > >SH> Create the following simple database containing just one BDB table: > >SH> CREATE DATABASE test; > >SH> CREATE TABLE foo ( >SH> id INT NOT NULL AUTO_INCREMENT, >SH> num INT, >SH> str VARCHAR(10), >SH> PRIMARY KEY (id) >SH> ) TYPE=BDB; > >SH> Now start two "mysql.exe" sessions running with the "-f" flag to ignore >SH> errors. > >SH> Set each one into "no auto-commit" mode with: > >SH> SET AUTOCOMMIT=0; > >SH> and then feed each of them the following two lines 50,000 times each: > >SH> INSERT INTO foo (num, str) VALUES (1, 'a'); >SH> COMMIT; > >[skip] > >SH> When I inspect the output collected in these log files I find, not >SH> unexpectedly, that there are many "deadlock" errors (which I'm quite >SH> comfortable with, and can handle in my real software by simply retrying >SH> the transaction), but that there are also a few of the following error: > >SH> Duplicate entry '1' for key 1 > >SH> Surely this is a bug isn't it? > >Steve, I tested your example with PHP script, but I didn't get any >Duplicate entry error ... >(I tested with BDB tables and SET AUTOCOMMIT=0). >
What was the PHP script that you tested it with? Perhaps I could try the same PHP script myself and see what I get. I've also found that my test fails on some machines and works on others; sometimes it runs on a lot further than 50,000 iterations on some machines before it fails. My desktop machine is a WinNT4 PIII 667MHz. What sort of machine have you tried on? - Steve --------------------------------------------------------------------- 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
