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

> On 24 May 2022, at 14:02, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
>
> I approached it yesterday by running the test case with each
> set_indexsafe_procflags() callsite commented out, see which one breaks
> things

On my machine commenting out set_indexsafe_procflags() before "Phase 3 of concurrent index build" fixes the tests.

> On 23 May 2022, at 23:18, Peter Geoghegan <pg(at)bowt(dot)ie> wrote:
>
> Maybe you'd get a clearer/earlier failure if you also applied this
> patch, on an assertion-enabled build

I've tried this approach, but nothing actually seem to change... BTW I used a three-way-merge rebase, there is a slight conflict in comments. But, luckily, comments don't run.

> On 24 May 2022, at 06:20, Michael Paquier <michael(at)paquier(dot)xyz> wrote:
>
> Thanks for the TAP test. That's nice. It actually passes here,
> reliably.

IDK. Maybe if you increase --time of pgbench you will observe the problem...

> On 24 May 2022, at 07:19, Peter Geoghegan <pg(at)bowt(dot)ie> wrote:
>
> Andrey was just
> making an observation, in case it helped with debugging.

Yes, I'm not proposing to commit anything so far. All my tests, snippets, diffs here are only debug stuff.

Thank you!

Best regards, Andrey Borodin.