Ah, well, that is just a SQL script you know. :)
On 12/28/05, Ken Payton wrote:
>
> Sounds correct. The problem seems to be with how statspack is reporting
> Rollback per transaction %. Apparently, from Allen's discovery, it
> seems statspack is using 'user rollbacks' as opposed to 'transaction
> rollbacks' in it's calculation. It can be slightly alarming when in the
> situation we were discussing earlier when working with connection
> pools. Oracle should probably change their conversion to use
> 'transaction rollbacks', in the meantime it makes better sense for us to
> perform our own calculation.
>
--
Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist
Received on Fri Dec 30 2005 - 19:05:32 CST
