MIME-Version: 1.0
References: <CA+TgmobSc_x6thvXZvHoni5Gs5-wsxyTRiOMKoeuX5br0PCtDA@mail.gmail.com>
 <CACX+KaPv4apqG3=Ef+FB9nn4C4cd6Z+604ej0PPOHKExH45u2A@mail.gmail.com>
 <934709.1739829723@sss.pgh.pa.us> <CAKFQuwZThU_Z-Zw+3mr+ecp1BVOw777dp3nXU5-wTVk3kS10gw@mail.gmail.com>
 <CA+TgmoZG71zBpLOfCGZqGhtp=88z6=YYhi54TEsCtKr3v+UpoA@mail.gmail.com>
 <1243984.1739894558@sss.pgh.pa.us> <CA+Tgmob_W0iq9Kuugra3WYTO2429RMJ_+HkVukrXWOUN81QiEw@mail.gmail.com>
 <1246906.1739896202@sss.pgh.pa.us> <CA+TgmoZYM2az+yCWu5DBnV50N_BE9f1r8-Doy6-tZTySeb-s+A@mail.gmail.com>
 <CACX+KaNAbOzePn710EtzH9F5xiUdBC+u59=UMab=Wr8jgDKQtw@mail.gmail.com> <Z8dcGMMP3-D5dobY@nathan>
In-Reply-To: <Z8dcGMMP3-D5dobY@nathan>
From: Ayush Vatsa <ayushvatsa1810@gmail.com>
Date: Sat, 8 Mar 2025 20:34:40 +0530
Message-ID: <CACX+KaO4R9QDxbPSxSB0jNXFsqA6Jf=UPS+tyUvT_YvuP_grVA@mail.gmail.com>
Subject: Re: Clarification on Role Access Rights to Table Indexes
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: Robert Haas <robertmhaas@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, 
	"David G. Johnston" <david.g.johnston@gmail.com>, 
	PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Content-Type: multipart/alternative; boundary="0000000000009d2eee062fd60f46"
Archived-At: <https://www.postgresql.org/message-id/CACX%2BKaO4R9QDxbPSxSB0jNXFsqA6Jf%3DUPS%2BtyUvT_YvuP_grVA%40mail.gmail.com>
Precedence: bulk

--0000000000009d2eee062fd60f46
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

> I'm wondering whether setting missing_ok to true is correct here.  IIUC w=
e
> should have an AccessShareLock on the index, but I don't know if that's
> enough protection.

Since we are already opening the relation with rel =3D relation_open(relOid=
,
AccessShareLock);,
if relOid does not exist, it will throw an error. If it does exist, we
acquire an AccessShareLock,
preventing it from being dropped.

By the time we reach IndexGetRelation(), we can be confident that relOid
exists and is
protected by the lock. Given this, it makes sense to keep missing_ok =3D fa=
lse
here.

Let me know if you agree or if you see any scenario where
missing_ok =3D true would be preferable=E2=80=94I can update the condition
accordingly.

Thanks!
Ayush Vatsa
SDE AWS

--0000000000009d2eee062fd60f46
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">&gt; I&#39;m wondering whether setting missing_ok to true =
is correct here.=C2=A0 IIUC we<br>&gt; should have an AccessShareLock on th=
e index, but I don&#39;t know if that&#39;s<br>&gt; enough protection.=C2=
=A0<div><p><font face=3D"arial, sans-serif">Since we are already opening th=
e relation with <code style=3D"">rel =3D relation_open(relOid, AccessShareL=
ock);</code>, <br>if <code style=3D"">relOid</code> does not exist, it will=
 throw an error. If it does exist, we acquire an <code style=3D"">AccessSha=
reLock</code>, <br>preventing it from being dropped.</font></p><p><font fac=
e=3D"arial, sans-serif">By the time we reach <code>IndexGetRelation()</code=
>, we can be confident that <code>relOid</code> exists and is <br>protected=
 by the lock. Given this, it makes sense to keep <code>missing_ok =3D false=
</code> here.</font></p><p><font face=3D"arial, sans-serif">Let me know if =
you agree or if you see any scenario where <br><code>missing_ok =3D true</c=
ode> would be preferable=E2=80=94I can update the condition accordingly.</f=
ont></p><p><font face=3D"arial, sans-serif">Thanks!<br>Ayush Vatsa<br>SDE A=
WS</font></p></div></div>

--0000000000009d2eee062fd60f46--