> Innodb is not row-level locking -- it's memory-page-level-locking. A
> memory page is usually small, so it's almost like row-level locking,
> but not quite. Perhaps you're running up against that?
>
> What does the query log say which queries were running? How do you
> know it's the delete tha
Innodb is indeed row level locking. You are likely thinking of BDB
which uses memory page level locking.
gw
sheeri kritzer wrote:
> Innodb is not row-level locking -- it's memory-page-level-locking. A
> memory page is usually small, so it's almost like row-level locking,
> but not quite. Perha
om: sheeri kritzer [mailto:[EMAIL PROTECTED]
> > Sent: Friday, February 10, 2006 9:52 AM
> > To: Ady Wicaksono
> > Cc: mysql@lists.mysql.com
> > Subject: Re: Innodb table locked on delete
> >
> > Innodb is not row-level locking -- it's
> > memory-page-l
riday, February 10, 2006 9:52 AM
> To: Ady Wicaksono
> Cc: mysql@lists.mysql.com
> Subject: Re: Innodb table locked on delete
>
> Innodb is not row-level locking -- it's
> memory-page-level-locking. A memory page is usually small,
> so it's almost like row-level locki
Innodb is not row-level locking -- it's memory-page-level-locking. A
memory page is usually small, so it's almost like row-level locking,
but not quite. Perhaps you're running up against that?
What does the query log say which queries were running? How do you
know it's the delete that's taking
So where's the row locking?
I configure my database with Innodb + READ COMMITED, by this configuration
by using autocommit=1, delete should be done on data commited to disk.
Other thread should be able to insert/update.
CMIIW
> Tables are locked on delete because, like an update, they are chang
Tables are locked on delete because, like an update, they are changing data.
Imagine issuing a REPLACE statement after a DELETE statement. If the
DELETE locks the table, then the REPLACE happens AFTER the DELETE, as
you wanted. If the DELETE does not lock the table, then it's possible
the REPLAC
