AW: Reduce latch row cache objects with event 10089
Date: Wed, 18 Mar 2015 12:23:58 +0000
Message-ID: <09011014EB621E4CBC2536B62A1B6407C88B2F30_at_smxc002.trivadis.com>
Hallo Timur,
for given index I usually list from shared pool all statements/plans which use the index, their benchmarks ( CPU/IO , number of executions).
Then I check how effective are the execution plans, what are the access and filter predicates.
That is the main source , but in the same way I look in AWR which statements/plans used the index in the past and what were the benchmarks.
Best Regards,
Petr
Von: Timur Akhmadeev [timur.akhmadeev_at_gmail.com] Gesendet: Mittwoch, 18. März 2015 12:16 An: Petr Novak
Cc: jonathan_at_jlcomp.demon.co.uk; oracle-l_at_freelists.org Betreff: Re: Reduce latch row cache objects with event 10089
On Wed, Mar 18, 2015 at 9:08 AM, Petr Novak <Petr.Novak_at_trivadis.com<mailto:Petr.Novak_at_trivadis.com>> wrote: I check continually index usage in shared pool and AWR, there are only 1-2 index candidates for elimination.
Do not need to look into AWR for that. Start with listing indexes & indexed columns instead, just looking at what is indexed + at avg_data_blocks_per_key. Very often people introduce redundant & unnecessary/inefficient indexes which could be logically eliminated.
--
Regards
Timur Akhmadeev
--
http://www.freelists.org/webpage/oracle-l
Received on Wed Mar 18 2015 - 13:23:58 CET