MIME-Version: 1.0
References: 
 <CACG=ezaWg7_nt-8ey4aKv2w9LcuLthHknwCawmBgEeTnJrJTcw@mail.gmail.com>
 <CACG=ezaLdhpFY9_qr7B6s7kkg_=s5S_ZD2=dsSeBNgdWWWuKbg@mail.gmail.com>
 <CACG=ezaFoCx4XEGi4gQWrusYD81LTLTKBUdgs-mEjNwzVarRnw@mail.gmail.com>
 <CACG=ezarNdLaTdr3fdWGyVJEzRAa=Vd5LLmu7VN2r7XL6LH8xA@mail.gmail.com>
 <CAPpHfdtPybyMYBj-x3-Z5=4bj_vhYk2R0nezfy=Vjcz4QBMDgw@mail.gmail.com>
 <CACG=ezaCc4bFfua-VA1NB6wppMPwuMmZiGdrkb-iYK9ZmQa6gg@mail.gmail.com>
 <CAGjGUAJUmSFMunCcK8DXcjLrs2Hfk2kFiaWDTc6ti03S8Echmw@mail.gmail.com>
 <CACG=ezaCKR+--O2TZrm3jYGgpDjdpLXdz3marG6t_=nzP5+Gog@mail.gmail.com>
 <CAGjGUA+uxpbkRaxkZmcSNiGXJ_G3Zj4-gzaiy=94DH7rvE8tig@mail.gmail.com>
 <CACG=ezbPUASDL1eJ+c-ZkJMwRPukvp3EL0q1vSUa1h+fnX8y3g@mail.gmail.com>
 <fecc2b80-abe4-4fb7-9df9-3bbf54166eaa@iki.fi>
 <CACG=ezaABYDepYf24MUNxc2oHRERxXbXHNMP+i-Pr1AXu26x0A@mail.gmail.com>
 <0794857c-aaaf-4cd3-bd99-84c6155bb2f6@iki.fi>
 <CACG=ezb53xpK39y2oe8H5BAAEnrtpqAjTF_xKoFZh8z6PJY7zA@mail.gmail.com>
 <bc7a77a9-4158-418e-8aa7-788e926c96d9@iki.fi>
 <CACG=ezbZo_3_fnx=S5BfepwRftzrpJ+7WET4EkTU6wnjDTsnjg@mail.gmail.com>
In-Reply-To: 
 <CACG=ezbZo_3_fnx=S5BfepwRftzrpJ+7WET4EkTU6wnjDTsnjg@mail.gmail.com>
From: Maxim Orlov <orlovmg@gmail.com>
Date: Fri, 7 Nov 2025 19:03:11 +0300
Message-ID: 
 <CACG=ezYiEB8wVMuHVOnsA4Hc+BPG9Zo8Dz_G-EC6re-7P3Pb9A@mail.gmail.com>
Subject: Re: POC: make mxidoff 64 bits
To: Heikki Linnakangas <hlinnaka@iki.fi>
Cc: wenhui qiu <qiuwenhuifx@gmail.com>,
 Alexander Korotkov <aekorotkov@gmail.com>,
	Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>,
	Postgres hackers <pgsql-hackers@lists.postgresql.org>
Content-Type: multipart/alternative; boundary="0000000000002f23160643035212"
Archived-At: 
 <https://www.postgresql.org/message-id/CACG%3DezYiEB8wVMuHVOnsA4Hc%2BBPG9Zo8Dz_G-EC6re-7P3Pb9A%40mail.gmail.com>
Precedence: bulk

--0000000000002f23160643035212
Content-Type: text/plain; charset="UTF-8"

I noticed one minor issue after I had already sent the
previous letter.

--- a/src/backend/access/transam/multixact.c
+++ b/src/backend/access/transam/multixact.c
@@ -1034,7 +1034,7 @@ GetNewMultiXactId(int nmembers, MultiXactOffset
*offset)
    if (nextOffset + nmembers < nextOffset)
        ereport(ERROR,
                (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
-                "MultiXact members would wrap around"));
+                errmsg("MultiXact members would wrap around")));
    *offset = nextOffset;


$ $PGBINOLD/pg_controldata -D pgdata
pg_control version number:            1800
Catalog version number:               202510221
...
Latest checkpoint's NextMultiXactId:  10000000
Latest checkpoint's NextMultiOffset:  999995050
Latest checkpoint's oldestXID:        748
...

I tried finding out how long it would take to convert a big number of
segments. Unfortunately, I only have access to a very old machine right
now. It took me 7 hours to generate this much data on my old
Intel(R) Core(TM) i5-6500 CPU @ 3.20GHz with 16 Gb of RAM.

Here are my rough measurements:

HDD
$ sudo sync && echo 3 | sudo tee /proc/sys/vm/drop_caches
$ time pg_upgrade
...
real    4m59.459s
user    0m19.974s
sys     0m13.640s

SSD
$ sudo sync && echo 3 | sudo tee /proc/sys/vm/drop_caches
$ time pg_upgrade
...
real    4m52.958s
user    0m19.826s
sys     0m13.624s

I aim to get access to more modern stuff and check it all out there.

-- 
Best regards,
Maxim Orlov.

--0000000000002f23160643035212
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>I noticed one minor issue after I had already sent th=
e</div><div>previous letter.</div><div><br></div><div>--- a/src/backend/acc=
ess/transam/multixact.c<br>+++ b/src/backend/access/transam/multixact.c<br>=
@@ -1034,7 +1034,7 @@ GetNewMultiXactId(int nmembers, MultiXactOffset *offs=
et)<br>=C2=A0 =C2=A0 if (nextOffset + nmembers &lt; nextOffset)<br>=C2=A0 =
=C2=A0 =C2=A0 =C2=A0 ereport(ERROR,<br>=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),<br>- =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0&quot;MultiXact members wou=
ld wrap around&quot;));<br>+ =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0 =C2=A0errmsg(&quot;MultiXact members would wrap around&quot;)));<br>=C2=
=A0 =C2=A0 *offset =3D nextOffset;</div><div><br></div><div><br></div><div>=
$ $PGBINOLD/pg_controldata -D pgdata<br>pg_control version number: =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A01800<br>Catalog version number: =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 202510221<br>...<br>Latest checkp=
oint&#39;s NextMultiXactId: =C2=A010000000<br>Latest checkpoint&#39;s NextM=
ultiOffset: =C2=A0999995050<br>Latest checkpoint&#39;s oldestXID: =C2=A0 =
=C2=A0 =C2=A0 =C2=A0748<br>...</div><div><br></div><div>I tried finding out=
 how long it would take to convert a big number of<br>segments. Unfortunate=
ly, I only have access to a very old machine right<br>now. It took me 7 hou=
rs to generate this much data on my old<br>Intel(R) Core(TM) i5-6500 CPU @ =
3.20GHz with 16 Gb of RAM.<br><br>Here are my rough measurements:<br><br></=
div><div>HDD<br>$ sudo sync &amp;&amp; echo 3 | sudo tee /proc/sys/vm/drop_=
caches<br>$ time pg_upgrade<br>...<br>real =C2=A0 =C2=A04m59.459s<br>user =
=C2=A0 =C2=A00m19.974s<br>sys =C2=A0 =C2=A0 0m13.640s<br><br>SSD<br>$ sudo =
sync &amp;&amp; echo 3 | sudo tee /proc/sys/vm/drop_caches<br>$ time pg_upg=
rade<br>...<br>real =C2=A0 =C2=A04m52.958s<br>user =C2=A0 =C2=A00m19.826s<b=
r>sys =C2=A0 =C2=A0 0m13.624s</div><div><br></div><div>I aim to get access =
to more modern stuff and check it all out there.</div><div><br></div><span =
class=3D"gmail_signature_prefix">-- </span><br><div dir=3D"ltr" class=3D"gm=
ail_signature"><div dir=3D"ltr"><div>Best regards,<br></div><div>Maxim Orlo=
v.</div></div></div></div>

--0000000000002f23160643035212--