MIME-Version: 1.0
References: <CAB+Nuk8y88kvoRHsfN2Pjed9o_iM+9FbveU=4P8G0EQbzFP+qw@mail.gmail.com>
 <15AB8DDB-7D4B-4272-801F-F5DD84E829E6@thebuild.com> <DDF016B5-F5CD-4143-A2C7-89BDE6F61FBE@yesql.se>
In-Reply-To: <DDF016B5-F5CD-4143-A2C7-89BDE6F61FBE@yesql.se>
From: Ron Johnson <ronljohnsonjr@gmail.com>
Date: Wed, 7 Aug 2024 09:52:49 -0400
Message-ID: <CANzqJaCXgvqqgev=SKaQvLu9RwjDfR_KXo_=Xpffm=jBkunvPg@mail.gmail.com>
Subject: Re: data checksums
To: Daniel Gustafsson <daniel@yesql.se>
Cc: Christophe Pettus <xof@thebuild.com>, bruno vieira da silva <brunogiovs@gmail.com>, 
	pgsql-general@lists.postgresql.org
Content-Type: multipart/alternative; boundary="000000000000724b2c061f183a9a"
Archived-At: <https://www.postgresql.org/message-id/CANzqJaCXgvqqgev%3DSKaQvLu9RwjDfR_KXo_%3DXpffm%3DjBkunvPg%40mail.gmail.com>
Precedence: bulk

--000000000000724b2c061f183a9a
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, Aug 7, 2024 at 3:41=E2=80=AFAM Daniel Gustafsson <daniel@yesql.se> =
wrote:

> > On 6 Aug 2024, at 18:29, Christophe Pettus <xof@thebuild.com> wrote:
> >> On Aug 6, 2024, at 08:11, bruno vieira da silva <brunogiovs@gmail.com>
> wrote:
>
> >> the pg doc
> >> mentions a considerable performance penality, how considerable it is?
> >
> > That line is probably somewhat out of date at this point.  We haven't
> seen a significant slowdown in enabling them on any modern hardware.  I
> always turn them on, except on the type of filesystems/hardware mentioned
> above.
>
> The last in-depth analysis of data checksums (and hint bits) overhead tha=
t
> I
> can remember is from 2019:
>
>
> https://www.postgresql.org/message-id/20190330192543.GH4719%40development


A quote from that post:
"I have not investigated the exact reasons, but my hypothesis it's about
the amount of WAL generated during the initial CREATE INDEX (because it
probably ends up setting the hint bits), which puts additional pressure
on the storage."

Presuming that hypothesis is true: how often do "you" run CREATE INDEX (or
VACUUM FULL or CLUSTER)?  I certainly don't run them very often.

--=20
Death to America, and butter sauce!
Iraq lobster...

--000000000000724b2c061f183a9a
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr">On Wed, Aug 7, 2024 at 3:41=E2=80=AFAM Da=
niel Gustafsson &lt;<a href=3D"mailto:daniel@yesql.se">daniel@yesql.se</a>&=
gt; wrote:<br></div><div class=3D"gmail_quote"><blockquote class=3D"gmail_q=
uote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,2=
04);padding-left:1ex">&gt; On 6 Aug 2024, at 18:29, Christophe Pettus &lt;<=
a href=3D"mailto:xof@thebuild.com" target=3D"_blank">xof@thebuild.com</a>&g=
t; wrote:<br>
&gt;&gt; On Aug 6, 2024, at 08:11, bruno vieira da silva &lt;<a href=3D"mai=
lto:brunogiovs@gmail.com" target=3D"_blank">brunogiovs@gmail.com</a>&gt; wr=
ote:<br>
<br>
&gt;&gt; the pg doc <br>
&gt;&gt; mentions a considerable performance penality, how considerable it =
is?<br>
&gt; <br>
&gt; That line is probably somewhat out of date at this point.=C2=A0 We hav=
en&#39;t seen a significant slowdown in enabling them on any modern hardwar=
e.=C2=A0 I always turn them on, except on the type of filesystems/hardware =
mentioned above.<br>
<br>
The last in-depth analysis of data checksums (and hint bits) overhead that =
I<br>
can remember is from 2019:<br>
<br>
=C2=A0 <a href=3D"https://www.postgresql.org/message-id/20190330192543.GH47=
19%40development" rel=3D"noreferrer" target=3D"_blank">https://www.postgres=
ql.org/message-id/20190330192543.GH4719%40development</a></blockquote><div>=
<br></div><div>A quote from that post:</div><div>&quot;<span style=3D"color=
:rgb(0,0,0);font-family:monospace,monospace;font-size:10.7333px">I have not=
 investigated the exact reasons, but my hypothesis it&#39;s about</span></d=
iv><div><span style=3D"color:rgb(0,0,0);font-family:monospace,monospace;fon=
t-size:10.7333px">the amount of WAL generated during the initial CREATE IND=
EX (because it</span><br style=3D"box-sizing:border-box;color:rgb(0,0,0);fo=
nt-family:monospace,monospace;font-size:10.7333px"><span style=3D"color:rgb=
(0,0,0);font-family:monospace,monospace;font-size:10.7333px">probably ends =
up setting the hint bits), which puts additional pressure</span><br style=
=3D"box-sizing:border-box;color:rgb(0,0,0);font-family:monospace,monospace;=
font-size:10.7333px"><span style=3D"color:rgb(0,0,0);font-family:monospace,=
monospace;font-size:10.7333px">on the storage.</span>&quot;</div><div>=C2=
=A0</div><div>Presuming that hypothesis is true: how often do &quot;you&quo=
t; run CREATE INDEX (or VACUUM FULL or CLUSTER)?=C2=A0 I certainly don&#39;=
t run them very often.</div></div><div><br></div><span class=3D"gmail_signa=
ture_prefix">-- </span><br><div dir=3D"ltr" class=3D"gmail_signature"><div =
dir=3D"ltr">Death to America, and butter sauce!<div>Iraq lobster...</div></=
div></div></div>

--000000000000724b2c061f183a9a--