Re: Interpreting utlbstat/utlestat output

Hi Tony,

here's something from Corey, Abbey, Dechichio's 'Tuning Oracle': The GET_REQS column shows how many times a particular type was requested. GET_MISS shows how often the request could not be solved from cache. Oracle then has to fetch this from disc. GET_MISS, SCAN_MISS should be below 10%. Increase SHARED_POOL_SIZE in you INIT<SID>.ORA to increase cache but keep in mind, that this will also increase the total amount of memory oracle uses.

Tony Mizukami wrote:
>
> Greetings.
>
> I just finished running the utlbstat/utlestat scripts, and came across
> the folowing in the output:
>
> ------------------------------------------------------------------------------
> SQLWKS> Rem get_miss and scan_miss should be very low compared to the
> requests.
> SQLWKS> Rem cur_usage is the number of entries in the cache that are
> being used.
> SQLWKS> select * from stats$dc
> 2> where get_reqs != 0 or scan_reqs != 0 or mod_reqs != 0;
> NAME GET_REQS GET_MISS SCAN_REQ SCAN_MIS MOD_REQS COUNT
> CUR_USAG
> --------------- -------- -------- -------- -------- -------- --------
> --------
> dc_tablespaces 2082 0 0 0 0
> 8 6
> dc_free_extents 9471 2933 1484 0 7322
> 77 36
> dc_segments 4556 199 0 0 1681
> 112 105
> dc_rollback_seg 1400 0 0 0 0
> 44 36
> dc_used_extents 2968 1484 0 0 2968
> 94 19
> .
> .
> .
> ------------------------------------------------------------------------------
>
> The remarks say that the get_miss and scan_miss values for these
> variables should be low compared to the requests, but my dc_free_extents
> and dc_used_extents values seem to be high.
>
> Can anyone explain what these mean, and how they can be fixed?
>
> Much thanks in advance.
>
> Tony Mizukami
> mizukami @ foxproductions . com
 

-- 
Regards

Matthias Gresz    :-)