Bug #21938 Race condition in update_auto_increment()

• Submitted: 31 Aug 2006 5:37
• Modified: 4 Nov 2008 23:18
• Reporter: Vadim Tkachenko
• Status: No Feedback
• Category: MySQL Server
• Severity: S4 (Feature request)
• Version: 5.1
• OS: Any
• CPU Architecture: Any

[31 Aug 2006 5:37] Vadim Tkachenko

Description:
This bug is related to http://bugs.mysql.com/16979 
where I proposed patch for InnoDB/auto_increment.

The problem handler function
bool handler::update_auto_increment()
is not protected from concurrent calls and storage engine must use its own locks
to get consistent values that tends to performance degradation.

How to repeat:
look in to source code

Suggested fix:
update_auto_increment() should be protected without external locks

[26 Sep 2006 14:46] Calvin Sun

change it to feature request.

[28 Sep 2006 17:14] Trudy Pelzer

Per Sinisa, this is just a 5.1 problem, but
affects all transactional storage engines.

[4 Oct 2008 23:18] Konstantin Osipov

Impossible to verify - no test case.
Work has been ongoing on update_auto_increment(). 
It's not clear what the complain is about.

[4 Oct 2008 23:18] Konstantin Osipov

(It's not clear what it is about looking at the current 5.1 code).

[5 Nov 2008 0:00] Bugs System

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".