MIME-Version: 1.0
References: <20240222.114600.1019580904613326727.horikyota.ntt@gmail.com>
 <CAHewXN=ZEL1Rvw2fJmt8k3sO_3fb58gijv7GR-JnY8ujgGP0UA@mail.gmail.com>
 <20240222.143815.1856873109774774353.horikyota.ntt@gmail.com> <20240222.160027.62383325429446607.horikyota.ntt@gmail.com>
In-Reply-To: <20240222.160027.62383325429446607.horikyota.ntt@gmail.com>
From: Alex Masterov <amasterov@gmail.com>
Date: Wed, 6 May 2026 13:26:51 +0200
Message-ID: <CA+8z=zu-mxXuBtnTVNcOzaGJhgeKSTyeAsK+AdRMr5_XZM1SDw@mail.gmail.com>
Subject: Re: BUG #18354: Aborted transaction aborted during cleanup when
 temp_file_limit exceeded
To: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Cc: tndrwang@gmail.com, exclusion@gmail.com, pgsql-bugs@lists.postgresql.org
Content-Type: multipart/mixed; boundary="00000000000054e745065124719a"
Archived-At: <https://www.postgresql.org/message-id/CA%2B8z%3Dzu-mxXuBtnTVNcOzaGJhgeKSTyeAsK%2BAdRMr5_XZM1SDw%40mail.gmail.com>
Precedence: bulk

--00000000000054e745065124719a
Content-Type: multipart/alternative; boundary="00000000000054e7440651247198"

--00000000000054e7440651247198
Content-Type: text/plain; charset="UTF-8"

Hi,

 While running tests with Neon, we discovered an assertion failure that can
 occur during re-entrant AbortTransaction() calls.

 The issue arises when an error occurs during AbortTransaction() after
 ProcArrayEndTransaction() has cleared MyProc->xid. If another error is
raised
 during cleanup (e.g., in AtEOXact_Inval()), the PostgresMain error handler
 invokes AbortCurrentTransaction() again. The second AbortTransaction() call
 reads a still-valid s->transactionId (CleanupTransaction() hasn't run yet)
 and passes it to ProcArrayEndTransaction(), which then hits:

   Assert(TransactionIdIsValid(proc->xid))

 because MyProc->xid was already cleared by the first call.

 The attached patch fixes this by checking MyProc->xid validity before
calling
 RecordTransactionAbort() and only passing a valid latestXid when
appropriate.

 **Reproduction:**
 This can be reproduced reliably using the injection_points extension:

 1. Attach the injection point:
   SELECT injection_points_attach('transaction-end-process-inval', 'error');
 2. Create invalidation messages: CREATE TABLE test(id int);
 3. Trigger abort: ROLLBACK;

 Without the fix: assertion crash on ProcArrayEndTransaction()
 With the fix applied: the script will panic with "ERRORDATA_STACK_SIZE
exceeded"
 due to re-entrant error handling, demonstrating that the assertion is
resolved.

 I've included a reproduction script and the fix that clearly shows both
behaviors.

 **Files attached:**
 - 0001-xact-Prevent-assertion-failure-in-re-entrant-Abort.patch
 - repro_minimal_panic_if_fixed.sh

 Thoughts?

 Best regards,
 Alexey

>
>
>
>
>
>
>

--00000000000054e7440651247198
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr"><br></div>Hi,<br><br>=C2=A0While running =
tests with Neon, we discovered an assertion failure that can<br aria-hidden=
=3D"true">=C2=A0occur during re-entrant AbortTransaction() calls.<br><br>=
=C2=A0The issue arises when an error occurs during AbortTransaction() after=
<br aria-hidden=3D"true">=C2=A0ProcArrayEndTransaction() has cleared MyProc=
-&gt;xid. If another error is raised<br aria-hidden=3D"true">=C2=A0during c=
leanup (e.g., in AtEOXact_Inval()), the PostgresMain error handler<br aria-=
hidden=3D"true">=C2=A0invokes AbortCurrentTransaction() again. The second A=
bortTransaction() call<br aria-hidden=3D"true">=C2=A0reads a still-valid s-=
&gt;transactionId (CleanupTransaction() hasn&#39;t run yet)<br aria-hidden=
=3D"true">=C2=A0and passes it to ProcArrayEndTransaction(), which then hits=
:<br><br>=C2=A0=C2=A0=C2=A0Assert(TransactionIdIsValid(proc-&gt;xid))<br><b=
r>=C2=A0because MyProc-&gt;xid was already cleared by the first call.<br><b=
r>=C2=A0The attached patch fixes this by checking MyProc-&gt;xid validity b=
efore calling<br aria-hidden=3D"true">=C2=A0RecordTransactionAbort() and on=
ly passing a valid latestXid when appropriate.<br><br>=C2=A0**Reproduction:=
**<br aria-hidden=3D"true">=C2=A0This can be reproduced reliably using the =
injection_points extension:<br><br>=C2=A01. Attach the injection point:<br =
aria-hidden=3D"true">=C2=A0=C2=A0=C2=A0SELECT injection_points_attach(&#39;=
transaction-end-process-inval&#39;, &#39;error&#39;);<br aria-hidden=3D"tru=
e">=C2=A02. Create invalidation messages: CREATE TABLE test(id int);<br ari=
a-hidden=3D"true">=C2=A03. Trigger abort: ROLLBACK;<br><br>=C2=A0Without th=
e fix: assertion crash on ProcArrayEndTransaction()<br aria-hidden=3D"true"=
>=C2=A0With the fix applied: the script will panic with &quot;ERRORDATA_STA=
CK_SIZE exceeded&quot;<br aria-hidden=3D"true">=C2=A0due to re-entrant erro=
r handling, demonstrating that the assertion is resolved.<br><br>=C2=A0I=
9;ve included a reproduction script and the fix that clearly shows both beh=
aviors.<br><br>=C2=A0**Files attached:**<br aria-hidden=3D"true">=C2=A0- 00=
01-xact-Prevent-assertion-failure-in-re-entrant-Abort.patch<br aria-hidden=
=3D"true">=C2=A0- repro_minimal_panic_if_fixed.sh<br><br>=C2=A0Thoughts?<br=
><br>=C2=A0Best regards,<br aria-hidden=3D"true">=C2=A0Alexey<br><div class=
=3D"gmail_quote gmail_quote_container"><blockquote class=3D"gmail_quote" st=
yle=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padd=
ing-left:1ex"><br><br>
<br>
<br>
<br>
<br>
</blockquote></div></div>

--00000000000054e7440651247198--
--00000000000054e745065124719a
Content-Type: application/octet-stream; 
	name="0001-xact-Prevent-assertion-failure-in-re-entrant-Abort.patch"
Content-Disposition: attachment; 
	filename="0001-xact-Prevent-assertion-failure-in-re-entrant-Abort.patch"
Content-Transfer-Encoding: base64
Content-ID: <f_motz0v3d0>
X-Attachment-Id: f_motz0v3d0
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--00000000000054e745065124719a
Content-Type: text/x-sh; charset="UTF-8"; name="repro_minimal_panic_if_fixed.sh"
Content-Disposition: attachment; filename="repro_minimal_panic_if_fixed.sh"
Content-Transfer-Encoding: base64
Content-ID: <f_motz159e1>
X-Attachment-Id: f_motz159e1
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--00000000000054e745065124719a--