Re: IRe: BUG #16792: silent corruption of GIN index resulting in SELECTs returning non-matching rows

On 14/07/2021 17:37, Pawel Kudzia wrote:
> Heikki - the data is semi-sensitive. I cannot share the whole table.
> I could if it was possible to remove some columns or isolate only
> handful of rows.
>
> But - just doing the "vacuum full entity" resolves the original issue
> and select returns 0 rows, as it should.

Ok. One idea would be to write a program or extension that overwrites
all the uninteresting columns with zeros.

> How can I identify relevant pages to make selective dump of them for
> you, using method proposed by Peter?

It's a bit hard without knowing which index pages are affected. But the
attached script (dump-gin-page-info.sql) dumps some summary information
about the kind of pages there are. That's a start.

> Would interactive session where you can get temporary access to the
> server with copy of the database,
> to run gdb on it be of any help?

Let's try one thing first: I rebased a patch to add GIN support to the
amcheck extension [1]. I kind of doubt that it will catch the corruption
in this case, but you never know. So please apply the attached
v2-0001-Amcheck-for-GIN-13stable.patch patch. It is the same patch I
posted at [1], but the installation script is slightly different to make
it work on PostgreSQL v13. Then compile install it:

cd contrib/amcheck; make install

Then in psql:

CREATE EXTENSION amcheck;
gin_index_parent_check('entity_attribute_name_ids_gin');

If it reports any inconsistencies, we can try to use pageinspect to
gather further evidence on them.

[1]
https://www.postgresql.org/message-id/a5bc2aad-464b-91bd-061d-28af0f9b634c%40iki.fi
- Heikki