Re: Please advise [Re: Restore locking: TRUNCATE resets AUTO_INCREMENT on InnoDB]

[Ingo Strüwing]

Hi Chuck,

Charles Bell, 28.07.2009 16:30:

...
> Ok, So I'm reading through this thread and I start to wonder ... since
> we're using InnoDB which uses the row-level insert mechanism of the
> default drivers, what do the rows in the backup image file contain for
> the auto increment column? I think it would be helpful to see a dump of
> a row to see if the auto increment value is in there.

Is this an action item for me? Extracting and translating a row from a
backup image file?

Note that I still don't know, how to proceed with WL#4844/Bug#42895 and
their dependencies Bug#40944 and Bug#41716. The patch that completes all
four is ready, but we don't restore the auto_increment value correctly.
This is not a "real" regression from the current code base, but the fix
might be understood as incomplete. (The current code uses to work if
RESTORE is not attacked by DML. Since this might be the normal case, a
regression can be claimed as the new code never restores auto_increment
for InnoDB.) (OTOH, clean restore of auto_increment was not a goal for
BACKUP/RESTORE so far.)

At first I need a decision, if the auto_increment problem must be fixed
by this patch. If not, there is hope to fix all the above #s before my
vacation. If yes, then I need a decision, how to fix it after my vacation.

Regards
Ingo
-- 
Ingo Strüwing, Database Group
Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten
Geschäftsführer: Thomas Schröder,   Wolfgang Engels,   Wolf Frenkel
Vorsitzender des Aufsichtsrates: Martin Häring   HRB München 161028