Re: Fix race condition in InvalidatePossiblyObsoleteSlot()

From: Bertrand Drouvot <bertranddrouvot(dot)pg(at)gmail(dot)com>
To: Michael Paquier <michael(at)paquier(dot)xyz>
Cc: Bharath Rupireddy <bharath(dot)rupireddyforpostgres(at)gmail(dot)com>, pgsql-hackers(at)lists(dot)postgresql(dot)org, exclusion(at)gmail(dot)com
Subject: Re: Fix race condition in InvalidatePossiblyObsoleteSlot()
Date: 2024-02-26 14:01:45
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers


On Tue, Feb 20, 2024 at 04:03:53PM +0000, Bertrand Drouvot wrote:
> Hi,
> On Tue, Feb 20, 2024 at 02:33:44PM +0900, Michael Paquier wrote:
> > On Tue, Feb 20, 2024 at 08:51:17AM +0900, Michael Paquier wrote:
> > > Prefixing these with "initial_" is fine, IMO. That shows the
> > > intention that these come from the slot's data before doing the
> > > termination. So I'm OK with what's been proposed in v3.
> >
> > I was looking at that a second time, and just concluded that this is
> > OK, so I've applied that down to 16, wordsmithing a bit the comments.
> Thanks!
> FWIW, I've started to write a POC regarding the test we mentioned up-thread.
> The POC test is based on what has been submitted by Michael in [1]. The POC test
> seems to work fine and it seems that nothing more is needed in [1] (at some point
> I thought I would need the ability to wake up multiple "wait" injection points
> in sequence but that was not necessary).
> I'll polish and propose my POC test once [1] is pushed (unless you're curious
> about it before).

Though [1] mentioned up-thread is not pushed yet, I'm Sharing the POC patch now
(see the attached).


Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services:

Attachment Content-Type Size
v1-0001-Add-regression-test-for-818fefd8fd.patch text/x-diff 6.1 KB

In response to


Browse pgsql-hackers by date

  From Date Subject
Next Message Tomas Vondra 2024-02-26 14:13:06 Re: Shared detoast Datum proposal
Previous Message Julien Rouhaud 2024-02-26 13:54:41 Re: RFC: Logging plan of the running query