Re: interpreting utlestat report

"Ed Stevens" <Ed.Stevens_at_nmm.nissan-usa.com> wrote in message news:8cl6rn$3un$1_at_nnrp1.deja.com...
> In reviewing utlestat reports on several of our databases I have
> noticed several things that raised my eybrows - not with the
> performance per se (though there is plenty there as well!) but with
> some things that just don't look "right."
>
> 1) negative numbers for "logons current"
> 2) negative numbers for "opened cursors current"
> 3) negative numbers for "session uga memory"
> 4) negative numbers for undo segment XACTS
> 5) counts for blocks_read and block_changes, but no physical reads or
> writes. I assume this indicates everything was in the buffer pool but
> am not ready to have my feet held to the fire on it.
>
> Also I've been playing around with the SHARED_POOL_SIZE, trying to get
> some get hit ratios back up above 90%. Perhaps I need more observation
> periods, but I see at one increase of the SHARED_POOL_SIZE, the get hit
> ratio for the SQL area improved but the table/procedure area got worse.
> In studying back through my materials from the tuning workshop, it
> seems that the only item to impact any of these is the
> SHARED_POOL_SIZE, which also impacts the data dictionary cache. How
> does Oracle8 determine how much of the shared pool to give to the
> different areas of the library cache and the data dictionary cache?
>
> TIA
>
> - Ed Stevens
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.

Are you eyebrows still moving?
1 - 4 This means the current number (when utlestat was run) is lower than the number when utlbstat was run. Both figures are simply substracted. You can see that in the utlestat code.
5 This is correct.

Hth,

Sybrand Bakker, Oracle DBA Received on Fri Apr 07 2000 - 00:00:00 CDT