Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> Re: Lookup/Code table as hot block

Re: Lookup/Code table as hot block

From: Ranko Mosic <ranko.mosic_at_gmail.com>
Date: Sat, 22 Apr 2006 18:25:30 -0400
Message-ID: <367369f10604221525s18a19cetad6a1232dcabcd9a@mail.gmail.com>


Thanks.

On 4/21/06, Alex Gorbachev <gorbyx_at_gmail.com> wrote:
>
> Consider reading these 3 rows in the beginning of the session. You can
> have consider invalidating the cache after certain time and relod it.
> This is suitable if you have persistent connections and data might
> change on the server from time to time.
>
> 2006/4/21, Ranko Mosic <ranko.mosic_at_gmail.com>:
> >
> > But then you have to manually keep application in sync with table - no
> good.
> > Current idea is ( on app side ) to read it once at the beggining of the
> > session, cache into array and use from there - so buffer cache is not
> used
> > any more, just UGA ( as Tanel suggested ).
> >
> > Regards, Ranko
> >
> >
> > On 4/21/06, Ranko Mosic <ranko.mosic_at_gmail.com> wrote:
> > >
> > >
> > > That's the idea actually. Those 3 rows should stay away from Oracle.
> > > Thanks Alex.
> > >
> > >
> > >
> > > On 4/21/06, Alex Gorbachev <gorbyx_at_gmail.com> wrote:
> > > > Why don't you cache 3 rows in the application? You would avoid all
> > > > kind of problems with concurrent access then.
> > > >
> > > > 2006/4/20, Ranko Mosic <ranko.mosic_at_gmail.com>:
> > > > >
> > > > > Thanks all for their responses. I apologize for not having more
> detail
> > -
> > > > > this is yet another
> > > > > 4 pm issue. Version is 9.2.0.x. This table has only 3 ( three )
> rows,
> > so any
> > > > > points relating to pk/indexes/IOT etc don't apply here.
> > > > > Why is this bad thing ? I don't know, I asked client the same
> > question, but
> > > > > they seem sure
> > > > > they are seeing problem. I didn't have chance yet to have direct
> look
> > at the
> > > > > problem.
> > > > > Caching is current idea and also suggested by Oracle Consulting.
> > > > > I am not sure how much it will help though - block is cached
> anyway,
> > so
> > > > > buffer cache latch will be replaced with shared pool latch.
> Spreading
> > > > > accross 3 blocks is also considered.
> > > >
> > > >
> > > > --
> > > > Best regards,
> > > > Alex Gorbachev
> > > >
> > >
> > >
> > >
> > >
> > > --
> > > Regards,
> > > Ranko Mosic
> > > Contract Senior Oracle DBA
> > > B. Eng, Oracle 10g, 9i Certified Database Professional
> > > Phone: 416-450-2785
> > > email: mosicr_at_rogers.com
> > >
> >
> http://ca.geocities.com/mosicr@rogers.com/ContractSeniorOracleDBARankoMosicMain.html
> > >
> >
> >
> >
> > --
> > Regards,
> > Ranko Mosic
> > Contract Senior Oracle DBA
> > B. Eng, Oracle 10g, 9i Certified Database Professional
> > Phone: 416-450-2785
> > email: mosicr_at_rogers.com
> >
> http://ca.geocities.com/mosicr@rogers.com/ContractSeniorOracleDBARankoMosicMain.html
> >
>
>
> --
> Best regards,
> Alex Gorbachev
>

--
Regards,
Ranko Mosic
Contract Senior Oracle DBA
B. Eng, Oracle 10g, 9i Certified Database Professional
Phone: 416-450-2785
email: mosicr_at_rogers.com
http://ca.geocities.com/mosicr@rogers.com/ContractSeniorOracleDBARankoMosicMain.html

--
http://www.freelists.org/webpage/oracle-l
Received on Sat Apr 22 2006 - 17:25:30 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US