Re: How to drop a plan_hash_value from shared pool
Date: Wed, 18 Sep 2019 06:53:48 +0100
Message-ID: <CACj1VR7a-a-na9PSNyfOUEP41DshwbbqV5cT0STdXo_Bz+HTxQ_at_mail.gmail.com>
On Wed, 18 Sep 2019 at 03:51, Andrew Kerber <andrew.kerber_at_gmail.com> wrote:
> one method would be alter system flush shared pool; That’s kind of
> extreme though. Gets the bath water and the baby as it were.
>
> Sent from my iPhone
>
> > On Sep 17, 2019, at 5:48 PM, Mladen Gogala <gogala.mladen_at_gmail.com>
> wrote:
> >
> >
> >> On 9/17/19 6:20 PM, Jonathan Lewis wrote:
> >> There's a quarantine feature in the 19c autonomous database that allows
> for "quarantine" of badly behaving statements which, I think, allows for a
> particular plan (i.e. plan_hash_value) for a statement to be flagged as
> unacceptable. I haven't looked closely at the feature, but I wonder if it
> is something that says "if PHV nnnn is generated for SQL_ID xxxxx, don't
> use it", or whether it can be set to behave like "force_match". Or maybe it
> would even allow a specific PHV to be blocked irrespective of the statement
> that produced it.
> >>
> >>
> >> Regards
> >> Jonathan Lewis
> >>
> > My understanding is that quarantine is only available on Oracle
> engineered systems, just like heat maps. I am aware of the parameter to
> turn on the Exadata features:
> >
> > alter system set “_exadata_feature_on”=true scope=spfile;
> >
> > However, I am tinkering with that only in my own lab because I don't
> want to inflict pain to my employer. This is another potentially very
> useful feature, just like heat maps, restricted to the Oracle engineered
> systems only. I am not sure whether that would even work on ODA which is a
> neglected child among Oracle engineered systems.
> >
> >
> > --
> > Mladen Gogala
> > Database Consultant
> > Tel: (347) 321-1217
> >
> > --
> > http://www.freelists.org/webpage/oracle-l
> >
> >
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>
-- http://www.freelists.org/webpage/oracle-lReceived on Wed Sep 18 2019 - 07:53:48 CEST