Re: DB File Sequential Read Waits

Daniel Morgan <damorgan_at_x.washington.edu> wrote in message news:<1072754248.554601_at_yasure>...
> rjpfal wrote:
>
> > norwoodthree_at_my-deja.com (NorwoodThree) wrote in message news:<ba03e2c.0312291455.64db1f2d_at_posting.google.com>...
> >
> >>I agree with the above post...an Oracle upgrade would be the best use
> >>of your time. There are significant perf improvements in 9i if you
> >>upgrade to 9i. It is ** possible ** the hardware tweaks you are
> >>making could be avoided entirely.
> >
> >
> > I appreciate the advice but I am stuck with the version of Oracle due
> > to application dependency. There is planned migration in the coming
> > year but need to address situation as it stands now.
>
> There is no such thing as application dependency.
>
> Get 8.1.7.4.

I guess I should have phrased it differently. I confirmed that the app can actually run on 8.1.7.4 this morning. However, they are not planning on moving to 8.1.7.4 presently because production is running on 8.1.6.3.

> Open the init.ora with a text editor
> Type the following:
>
> compatible=8.1.6.3
>
> There is no application I have ever encountered that could possibly know
> about or care about the difference. They do, however, run better. And if
> your support contract with Oracle is current the cost is zero.

Will have to work in switch to 8.1.7.4 but in the meantime I still need to address present situation. There is bad SQL (working in using binds) and such random access to data that is seems it is seldom being found in memory. Then there is the IO, no significant improvement on IO time with new array using striped volumes. Could 8.1.6.3 actually be a cause to not find blocks in mem and go to disk. I am going to look on metalink for patches this morning. If there is a case then I can have better rational to move to 8.1.7.4 sooner rather than later. Received on Tue Dec 30 2003 - 10:02:49 CST