Re: BUG #17949: Adding an index introduces serialisation anomalies.

On Mon, Jun 26, 2023 at 8:04 PM Heikki Linnakangas <hlinnaka(at)iki(dot)fi> wrote:
> Yes, +1 on the patches. Any chance of constructing test cases for these?

Thanks for looking. I can't think of any good way to test
deterministically. All three depend on unlucky scheduling. Perhaps
if we had one of those 'pause insertion' systems that I have heard
talk of.

> > /*
> > - * We only get here if the index is completely empty. Lock relation
> > - * because nothing finer to lock exists.
> > + * Since we have no pages locked, it's possible for another
> > + * transaction to insert data between _bt_search() and
> > + * PredicateLockRelation(). We have to try again after taking a
> > + * relation-level predicate lock, to close a narrow window where we
> > + * wouldn't scan concurrently inserted tuples, but the writer wouldn't
> > + * see our predicate lock.
> > */
>
> I'd like to keep the old comment here, it's good context, and add the
> new text in addition to the old.

Done.

> v3-0002-Fix-race-in-SSI-interaction-with-bitmap-heap-scan.patch: Can we
> keep the optimization when not using SSI?

Done.

I'll push these in a couple of days if there are no further comments.