Date: Thu, 9 Jul 2026 04:34:20 +0000
From: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: Kouber Saparev <kouber@gmail.com>, pgsql-hackers@lists.postgresql.org
Subject: Re: BF mamba failure
Message-ID: <ak8kzJypHwIPq3UF@bdtpg>
References: 
 <CAN4RuQs6ht4T7ph8NbxHZ+NmDmTrvczAxpyZmcqk+74tzsciuw@mail.gmail.com>
 <aMIJozPEP5OLI3Yj@paquier.xyz>
 <CAN4RuQv3OTJ1Owqn2U+7ZbvL-N-nZYAUjF_NG8o12ayFN-=0GQ@mail.gmail.com>
 <aMNrSgJt6_oRqh45@paquier.xyz>
 <CAN4RuQvQ3ATcYvfTR1LzJnUJXpo_F8mgz-+WxoZsyusLLmCwYA@mail.gmail.com>
 <aMor2AHvdbZWTwCD@paquier.xyz>
 <CAN4RuQv2tk6ELko_M8YC6eM1YgV1Z2r-Lht22b0ObCTPTTG5Aw@mail.gmail.com>
 <ak2S5046X1aaagr4@paquier.xyz>
 <ak4sO/IINIBT7y0J@bdtpg>
 <ak7TIsbToN2iMs4K@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <ak7TIsbToN2iMs4K@paquier.xyz>
Archived-At: <https://www.postgresql.org/message-id/ak8kzJypHwIPq3UF%40bdtpg>
Precedence: bulk

Hi,

On Thu, Jul 09, 2026 at 07:45:54AM +0900, Michael Paquier wrote:
> On Wed, Jul 08, 2026 at 10:53:47AM +0000, Bertrand Drouvot wrote:
> Due to the decision of using missing_ok=true in the end-of-transaction
> drops, that could indeed help.  I'd be curious to hear back after the
> next minor release.

Yeah, Kouber do you think you could upgrade to 17.11 or 18.5? (current schedule
is August 13th, 2026 ([1])).

> > Michael, I wonder if that would make sense to add this "set_next_oid" kind of
> > thing to a contrib module to test this kind of OID reuse issue. Same idea as
> > xid_wraparound.
> 
> That's probably larger than just the OIDs.  Maybe something that's
> able to do what pg_resetwal does for the control file could be useful.
> Publishing another footgun sounds a bit annoying, but we also have
> stuff like pg_surgery, that serve a similar purpose to shortcut the
> backend.

Yeah. In addition to the OIDs I could think about the mxid (but I think they would
be better suited in xid_wraparound). The others look way more difficult (if doable)
to do outside pg_resetwal.

So I'd be tempted to just add an oid_wraparound one and we could add this test
case too. Thoughts?

[1]: https://www.postgresql.org/developer/roadmap/

Regards,

-- 
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com