MIME-Version: 1.0
References: 
 <CACeKOO0ifCtNP6b=h-24EWnw_NJy-KhOc9mWOLZKAxkF_iEogw@mail.gmail.com>
 <20260323.141315.1197840295971477812.ishii@postgresql.org>
 <CACeKOO2fMX9f9hFVHCYEF9BUWV0xD21K2JPQFYzZSWVdC6ZrXQ@mail.gmail.com>
 <20260407.090802.1335576118037412011.ishii@postgresql.org>
In-Reply-To: <20260407.090802.1335576118037412011.ishii@postgresql.org>
From: Nadav Shatz <nadav@tailorbrands.com>
Date: Tue, 7 Apr 2026 07:45:03 +0200
Message-ID: 
 <CACeKOO0U6o1kXx9-Bm9y8hkd0f0nk_VtXCv3CX8+Z=aYHmGaLg@mail.gmail.com>
Subject: Re: Proposal: Recent mutated table tracking in memory
To: Tatsuo Ishii <ishii@postgresql.org>
Cc: pgpool-hackers@lists.postgresql.org
Content-Type: multipart/alternative; boundary="0000000000008ed958064ed849ca"
Archived-At: 
 <https://www.postgresql.org/message-id/CACeKOO0U6o1kXx9-Bm9y8hkd0f0nk_VtXCv3CX8%2BZ%3DaYHmGaLg%40mail.gmail.com>
Precedence: bulk

--0000000000008ed958064ed849ca
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi Tatsuo,

Yes I ran into it during the work on the feature. Let me know if you want
me to separately submit it.

Cheers

Nadav Shatz
Tailor Brands | CTO


On Tue, Apr 7, 2026 at 2:08=E2=80=AFAM Tatsuo Ishii <ishii@postgresql.org> =
wrote:

> Hi Nadav,
>
> > Hi Tatsuo,
> >
> > Thank you for the thorough review and the fix to the tests!. Here's the
> > updated patch addressing all your comments.
> >
> > re - replication_delay_source_cmd requirement
> >
> > Good catch =E2=80=95 the feature now also works when `delay_threshold_b=
y_time >
> 0`.
> > I've added the TTL update call to `check_replication_time_lag()` (the
> > pg_stat_replication path), not just
> > `check_replication_time_lag_with_cmd()`. The docs are updated to reflec=
t
> > that either `replication_delay_source_cmd` or `delay_threshold_by_time`
> can
> > provide the time-based delay.
> >
> > re - Documentation compile error
> >
> > Fixed =E2=80=95 the xref was pointing to `runtime-config-track-table-mu=
tation`
> but
> > the actual section ID is `runtime-config-table-mutation-map`.
> >
> > Thanks again and looking forward to hearing back from you.
>
> While looking into your patch, I noticed that following part would be
> better to be committed as a separate patch as it's actually a bug fix,
> not related to the feature: query cache is not invalidated if MERGE
> statement is executed. I will take care of it and get back to you.
>
> --- a/src/query_cache/pool_memqcache.c
> +++ b/src/query_cache/pool_memqcache.c
> @@ -1305,6 +1305,12 @@ pool_extract_table_oids(Node *node, int **oidsp)
>                 }
>                 return num_oids;
>         }
> +       else if (IsA(node, MergeStmt))
> +       {
> +               MergeStmt  *stmt =3D (MergeStmt *) node;
> +
> +               table =3D make_table_name_from_rangevar(stmt->relation);
> +       }
>         else if (IsA(node, ExplainStmt))
>         {
>                 ListCell   *cell;
>
> Thanks for letting me know about the bug!
>
> Regards,
> --
> Tatsuo Ishii
> SRA OSS K.K.
> English: http://www.sraoss.co.jp/index_en/
> Japanese:http://www.sraoss.co.jp
>

--0000000000008ed958064ed849ca
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"auto">Hi=C2=A0Tatsuo,</div><div dir=3D"auto"><br></div><div dir=
=3D"auto">Yes I ran into it during the work on the feature. Let me know if =
you want me to separately submit it.=C2=A0</div><div dir=3D"auto"><br></div=
><div dir=3D"auto">Cheers</div><div dir=3D"auto"><span style=3D"color:rgb(0=
,0,0)"><br></span></div><div dir=3D"auto"><span style=3D"color:rgb(0,0,0)">=
Nadav Shatz</span><br><div dir=3D"auto"><div dir=3D"rtl" class=3D"gmail_sig=
nature" data-smartmail=3D"gmail_signature"><div dir=3D"ltr"><div><div dir=
=3D"ltr"><div dir=3D"rtl"><div dir=3D"ltr"><font style=3D"color:rgb(0,0,0)"=
>Tailor Brands=C2=A0| CTO</font></div></div></div></div></div></div></div><=
/div><div><br></div><div><br><div class=3D"gmail_quote gmail_quote_containe=
r"><div dir=3D"ltr" class=3D"gmail_attr">On Tue, Apr 7, 2026 at 2:08=E2=80=
=AFAM Tatsuo Ishii &lt;<a href=3D"mailto:ishii@postgresql.org">ishii@postgr=
esql.org</a>&gt; wrote:<br></div><blockquote class=3D"gmail_quote" style=3D=
"margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;pad=
ding-left:1ex;border-left-color:rgb(204,204,204)">Hi Nadav,<br>
<br>
&gt; Hi Tatsuo,<br>
&gt; <br>
&gt; Thank you for the thorough review and the fix to the tests!. Here&#39;=
s the<br>
&gt; updated patch addressing all your comments.<br>
&gt; <br>
&gt; re - replication_delay_source_cmd requirement<br>
&gt; <br>
&gt; Good catch =E2=80=95 the feature now also works when `delay_threshold_=
by_time &gt; 0`.<br>
&gt; I&#39;ve added the TTL update call to `check_replication_time_lag()` (=
the<br>
&gt; pg_stat_replication path), not just<br>
&gt; `check_replication_time_lag_with_cmd()`. The docs are updated to refle=
ct<br>
&gt; that either `replication_delay_source_cmd` or `delay_threshold_by_time=
` can<br>
&gt; provide the time-based delay.<br>
&gt; <br>
&gt; re - Documentation compile error<br>
&gt; <br>
&gt; Fixed =E2=80=95 the xref was pointing to `runtime-config-track-table-m=
utation` but<br>
&gt; the actual section ID is `runtime-config-table-mutation-map`.<br>
&gt; <br>
&gt; Thanks again and looking forward to hearing back from you.<br>
<br>
While looking into your patch, I noticed that following part would be<br>
better to be committed as a separate patch as it&#39;s actually a bug fix,<=
br>
not related to the feature: query cache is not invalidated if MERGE<br>
statement is executed. I will take care of it and get back to you.<br>
<br>
--- a/src/query_cache/pool_memqcache.c<br>
+++ b/src/query_cache/pool_memqcache.c<br>
@@ -1305,6 +1305,12 @@ pool_extract_table_oids(Node *node, int **oidsp)<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 }<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 return num_oids;<br=
>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 }<br>
+=C2=A0 =C2=A0 =C2=A0 =C2=A0else if (IsA(node, MergeStmt))<br>
+=C2=A0 =C2=A0 =C2=A0 =C2=A0{<br>
+=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0MergeStmt=C2=A0 *st=
mt =3D (MergeStmt *) node;<br>
+<br>
+=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0table =3D make_tabl=
e_name_from_rangevar(stmt-&gt;relation);<br>
+=C2=A0 =C2=A0 =C2=A0 =C2=A0}<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 else if (IsA(node, ExplainStmt))<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 {<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 ListCell=C2=A0 =C2=
=A0*cell;<br>
<br>
Thanks for letting me know about the bug!<br>
<br>
Regards,<br>
--<br>
Tatsuo Ishii<br>
SRA OSS K.K.<br>
English: <a href=3D"http://www.sraoss.co.jp/index_en/" rel=3D"noreferrer" t=
arget=3D"_blank">http://www.sraoss.co.jp/index_en/</a><br>
Japanese:<a href=3D"http://www.sraoss.co.jp" rel=3D"noreferrer" target=3D"_=
blank">http://www.sraoss.co.jp</a><br>
</blockquote></div></div>

--0000000000008ed958064ed849ca--