MIME-Version: 1.0
References: <DM6PR11MB34831A86630FF3F2F247F070F4BC9@DM6PR11MB3483.namprd11.prod.outlook.com>
 <CAGtsp8=uWbyjyA9tS7jno9F1e8hOMO+S3Rp82FxbaGvzWZgY_Q@mail.gmail.com>
 <DB7PR10MB209040C214FB69FB0BD548D085BA9@DB7PR10MB2090.EURPRD10.PROD.OUTLOOK.COM>
 <CAORjz=PO_mAt+M3=P41HsLSZmN5N7nDURuBZ8_cufNCFe1mWLg@mail.gmail.com>
 <SA2PR10MB456903641FFA617B07DACFE4A3B69@SA2PR10MB4569.namprd10.prod.outlook.com>
 <cd686c7d-1d6b-bc99-a9d8-134bd97e9301@gmail.com> <DB7PR10MB2090BE6C93F3797B03FEA88C85B59@DB7PR10MB2090.EURPRD10.PROD.OUTLOOK.COM>
 <f9e39108-4e35-a8e3-5ac6-1dd6bbc8f073@gmail.com> <CAGtsp8kT8K6FN-8mXOfODgXAgR68tQ=ctN_ggw2zgkTUdBt8uw@mail.gmail.com>
 <trinity-64039e06-8341-4cfa-965c-0d0547f7afd4-1612289659462@3c-app-webde-bap50>
In-Reply-To: <trinity-64039e06-8341-4cfa-965c-0d0547f7afd4-1612289659462@3c-app-webde-bap50>
From: Franck Pachot <franck@pachot.net>
Date: Tue, 2 Feb 2021 20:21:18 +0100
Message-ID: <CAK6ito2nGZa5AYjUSNYqU9jp5wyXQg6J7RRt9vhJDnm4nRfHbw@mail.gmail.com>
Subject: Re: Re: [External] : Re: Question on gathering System Statistics
To: rogel@web.de
Cc: Jonathan Lewis <jlewisoracle@gmail.com>,
 "oracle-l@freelists.org" <oracle-l@freelists.org>
Content-Type: multipart/alternative; boundary="00000000000090386005ba5f60c1"
Sender: oracle-l-bounce@freelists.org
Errors-to: oracle-l-bounce@freelists.org
Precedence: normal
Reply-To: franck@pachot.net
--00000000000090386005ba5f60c1
Content-Type: text/plain; charset="UTF-8"

My usual recommendation for system statistics is: keep the defaults and if
there is a query that has a better execution plan with different values of
MREADTIM, SREADTIM, MBRC then the problem is not on system statistics but
physical data model. Because there's no good access path and the CBO tries
to find the "less worse one" with blind guesses on physical reads. Either
make full scan faster (like partitioning) or index access faster (like
covering index) and the problem is solved. The right access path should not
depend on block access milliseconds because nobody knows if, at execution
time, the I/O will be a cache hit in fast memory, or queued in front of a
busy disk with many sessions waiting for it.
Franck.

On Tue, Feb 2, 2021 at 7:14 PM <rogel@web.de> wrote:

>
> Jonathan,
>
> since no one here doubts your authority, you are surely right with that
> (per "argumentum ad verecundiam" ).
> However, my knowledge of Latin, which used to be quite profound in school
> for nine years tells me that you are wrong with that.
>
> Regards
> Matthias
>
> *Gesendet:* Dienstag, 02. Februar 2021 um 18:19 Uhr
> *Von:* "Jonathan Lewis" <jlewisoracle@gmail.com>
> *An:* "oracle-l@freelists.org" <oracle-l@freelists.org>
> *Betreff:* Re: [External] : Re: Question on gathering System Statistics
>
> Splitting hairs here, but:
>
> "argument from authority" is "argumentum ab auctoritate"   (e.g. "from
> year of experiance with the CBO I can assure you that gathering systems
> stats is not a good idea ... ").
>
> "argumentum ad verecundiam" is "appeal to authority". (e.g. Nigel Bayliss
> is the PM for the optimizer and he says you don't need to gather system
> stats ...).
>
>
> Regards
> Jonathan Lewis
>
>
> On Tue, 2 Feb 2021 at 15:33, Mladen Gogala <gogala.mladen@gmail.com>
> wrote:
>
>> I am one of those guys who has real problems with authority. When I was a
>> kid, I challenged my math teacher to prove the Pythagora theorem and I told
>> her that I don't believe it without proof. She had my parents come to
>> school. That nasty attitude was further reinforced by me getting a Bsc in
>> mathematics few years later. Latin name for the argument from authority is
>> "argumentum ad verecundiam" and is considered to be a logical error since
>> the Roman times. I need some kind of explanation why is that so. I have
>> nothing but respect for Maria, Nigel and Jonathan, but I need some kind of
>> explanation. BTW, my favorite movie of all times is "Life of Brian", which
>> also has something to do with challenging authority.
>>
> -- http://www.freelists.org/webpage/oracle-l

--00000000000090386005ba5f60c1
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>My usual recommendation for system statistics is: kee=
p the defaults and if there is a query that has a better execution plan wit=
h different values of MREADTIM, SREADTIM, MBRC then the problem is not on s=
ystem statistics but physical data model. Because there&#39;s no good acces=
s path and the CBO tries to find the &quot;less worse one&quot; with blind =
guesses on physical reads. Either make full scan faster (like partitioning)=
 or index access faster (like covering index) and the problem is solved. Th=
e right access path should not depend on block access milliseconds because =
nobody knows if, at execution time, the I/O will be a cache hit in fast mem=
ory, or queued in front of a busy disk with many sessions waiting for it.<b=
r></div><div>Franck.<br></div></div><br><div class=3D"gmail_quote"><div dir=
=3D"ltr" class=3D"gmail_attr">On Tue, Feb 2, 2021 at 7:14 PM &lt;<a href=3D=
"mailto:rogel@web.de">rogel@web.de</a>&gt; wrote:<br></div><blockquote clas=
s=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid r=
gb(204,204,204);padding-left:1ex"><div><div style=3D"font-family:Verdana;fo=
nt-size:12px"><div>=C2=A0</div>

<div>Jonathan,</div>

<div>=C2=A0</div>

<div>since no one here doubts your authority, you are surely right with tha=
t (per &quot;argumentum ad verecundiam&quot; <img alt=3D"" src=3D"https://c=
dn.webde.de/cdn/mail/client/wicket/resource/static-res/---/mc/img/smileys/d=
efault/s_01.gif" style=3D"margin: 4px 2px 0px; vertical-align: bottom;">).<=
/div>

<div>However, my knowledge of Latin, which used to be quite profound in sch=
ool for nine years tells me that you are wrong with that.</div>

<div>
<div>=C2=A0</div>

<div>Regards</div>

<div>Matthias</div>

<div>=C2=A0
<div name=3D"quote" style=3D"margin:10px 5px 5px 10px;padding:10px 0px 10px=
 10px;border-left:2px solid rgb(195,217,229)">
<div style=3D"margin:0px 0px 10px"><b>Gesendet:</b>=C2=A0Dienstag, 02. Febr=
uar 2021 um 18:19 Uhr<br>
<b>Von:</b>=C2=A0&quot;Jonathan Lewis&quot; &lt;<a href=3D"mailto:jlewisora=
cle@gmail.com" target=3D"_blank">jlewisoracle@gmail.com</a>&gt;<br>
<b>An:</b>=C2=A0&quot;<a href=3D"mailto:oracle-l@freelists.org" target=3D"_=
blank">oracle-l@freelists.org</a>&quot; &lt;<a href=3D"mailto:oracle-l@free=
lists.org" target=3D"_blank">oracle-l@freelists.org</a>&gt;<br>
<b>Betreff:</b>=C2=A0Re: [External] : Re: Question on gathering System Stat=
istics</div>

<div name=3D"quoted-content">
<div>
<div>
<div>=C2=A0</div>

<div>Splitting hairs here, but:</div>

<div>=C2=A0</div>

<div>&quot;argument from authority&quot; is &quot;argumentum ab auctoritate=
&quot;=C2=A0=C2=A0 (e.g. &quot;from year of experiance with the CBO I can a=
ssure you that gathering systems stats is not a good idea ... &quot;).</div=
>

<div>=C2=A0</div>

<div>&quot;argumentum ad verecundiam&quot; is &quot;appeal to authority&quo=
t;. (e.g. Nigel Bayliss is the PM for the optimizer and he says you don&#39=
;t need to gather system stats ...).</div>

<div>=C2=A0</div>

<div>=C2=A0</div>

<div>Regards</div>

<div>Jonathan Lewis</div>

<div>=C2=A0</div>
</div>
=C2=A0

<div class=3D"gmail_quote">
<div class=3D"gmail_attr">On Tue, 2 Feb 2021 at 15:33, Mladen Gogala &lt;<a=
 href=3D"mailto:gogala.mladen@gmail.com" target=3D"_blank">gogala.mladen@gm=
ail.com</a>&gt; wrote:</div>

<blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-=
left:1px solid rgb(204,204,204);padding-left:1ex">
<div>
<p>I am one of those guys who has real problems with authority. When I was =
a kid, I challenged my math teacher to prove the Pythagora theorem and I to=
ld her that I don&#39;t believe it without proof. She had my parents come t=
o school. That nasty attitude was further reinforced by me getting a Bsc in=
 mathematics few years later. Latin name for the argument from authority is=
 &quot;argumentum ad verecundiam&quot; and is considered to be a logical er=
ror since the Roman times. I need some kind of explanation why is that so. =
I have nothing but respect for Maria, Nigel and Jonathan, but I need some k=
ind of explanation. BTW, my favorite movie of all times is &quot;Life of Br=
ian&quot;, which also has something to do with challenging authority.</p>
</div>
</blockquote>
</div>
</div>
</div>
</div>
</div>
</div></div></div>
--
<a href=3D"http://www.freelists.org/webpage/oracle-l" target=3D"_blank">htt=
p://www.freelists.org/webpage/oracle-l</a>


</blockquote></div>

--00000000000090386005ba5f60c1--
--
http://www.freelists.org/webpage/oracle-l