Re: BUG #17485: Records missing from Primary Key index when doing REINDEX INDEX CONCURRENTLY

On Wed, May 25, 2022 at 10:32:43AM -0700, Peter Geoghegan wrote:
> On Wed, May 25, 2022 at 10:18 AM Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> > Right, I mean I'm not saying I *like* reverting, and I'm not disputing
> > it's a good feature. Just that, if we have to choose between this
> > feature and not having index corruption, we better not have index
> > corruption.... and I'm not seeing any way that we can just tweak this
> > and make it work.
>
> Are we any closer to deciding on a timeline, in light of recent discussion?
>
> I'm now convinced that an out-of-schedule release is probably the way
> to go. My thinking is: are we really going to make users wait for
> August 11th for a fix?

No, we are unlikely to wait until August.

What do we currently know?

* Caused by CREATE/REINDEX CONCURRENTLY
* No way to detect which indexes were created concurrently
* amcheck will be required to find corrupt indexes (options?)
* Heap is fine, only index is corrupt
* Bug since PG 14
* Not known why reports are only appearing now
* Feature will likely need to be removed rather than fixed

--
Bruce Momjian <bruce(at)momjian(dot)us> https://momjian.us
EDB https://enterprisedb.com

Indecision is a decision. Inaction is an action. Mark Batterson