Re: BUG #18396: Assert in gistFindCorrectParent() fails on inserting large tuples into gist index
Tender Wang <tndrwang@gmail.com>
From: Tender Wang <tndrwang@gmail.com>
To: Alexander Lakhin <exclusion@gmail.com>,
Heikki Linnakangas <hlinnaka@iki.fi>
Cc: pgsql-bugs@lists.postgresql.org
Date: 2024-03-20T06:22:28Z
Lists: pgsql-bugs
Alexander Lakhin <exclusion@gmail.com> 于2024年3月20日周三 13:00写道:
> Hello Heikki and Tender Wang,
>
> 19.03.2024 17:26, Heikki Linnakangas wrote:
> > On 19/03/2024 14:07, Tender Wang wrote:
> >> Thanks for your report. I can reproduce this issue.
> >> I try to delete the Assert, no coredump anymore.
> >> I need some time to learn GiST to find the root cause.
> >
> > If you're interested to work on this, Tender, maybe you can figure that
> out?
>
I will take some time to figure that out.
> >
>
> Thank you for looking at this!
>
> I'm also bothered by errors "no empty local buffer available" produced
> (on master, but not on REL_12_STABLE) when inserting larger tuples:
> INSERT INTO t SELECT '' FROM generate_series(1, 500) g;
> INSERT INTO t SELECT
> (SELECT string_agg('text' || g, ' ') FROM generate_series(1, 700) g)
> FROM generate_series(1, 30);
>
> Are those errors expected?
>
I think it is expected. More local buffers need here. I set temp_buffers
to 2GB.
this error didn't report. But stack depth limit exceeded. It is a hint that
to process
this case need more resource.
--
Tender Wang
OpenPie: https://en.openpie.com/
Commits
Same data as JSON:
GET /api/v1/messages/:b64id/commits
the thread's linked commits as JSON, with link sources.
API reference →
-
Relax assertion in finding correct GiST parent
- b92482dc3eff 13.21 landed
- a749c6f18fba 14.18 landed
- 3c0fe75c412b 15.13 landed
- 41932139882e 16.9 landed
- 6526d07948c5 17.5 landed
- 7afca7edef75 18.0 landed
-
Fix another bug in parent page splitting during GiST index build.
- 28d3c2ddcf91 17.0 cited