Re: WIP: Avoid creation of the free space map for small tables

From: John Naylor <jcnaylor(at)gmail(dot)com>
To: Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>
Cc: pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: WIP: Avoid creation of the free space map for small tables
Date: 2018-10-16 10:57:25
Message-ID: CAJVSVGVtumVV-oo1XrwZ_dS2VHfjGgyRs_641LUDuoqyCqemaw@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-hackers

On 10/15/18, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
> I think you can avoid calling RelationGetNumberOfBlocks, if you call
> smgrexists before

This is done in the attached v5, 0001.

> and for the purpose of vacuum, we can get that as an
> input parameter. I think one can argue for not changing the interface
> functions like RecordPageWithFreeSpace to avoid calling
> RelationGetNumberOfBlocks, but to me, it appears worth to save the
> additional system call.

This is done in 0002. I also added a check for the cached value of
pg_class.relpages, since it's cheap and may help non-VACUUM callers.

> [proposal for a cache of blocks to try]

That's interesting. I'll have to do some reading elsewhere in the
codebase, and then I'll follow up.

Thanks,
-John Naylor

Attachment Content-Type Size
v5-0001-Avoid-creation-of-the-free-space-map-for-small-ta.patch text/x-patch 17.5 KB
v5-0002-Add-parameter-nblocks-to-RecordPageWithFreeSpace.patch text/x-patch 10.6 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Amit Khandekar 2018-10-16 11:04:41 Re: TupleTableSlot abstraction
Previous Message Andrey Klychkov 2018-10-16 08:28:17 [PATCH] heap_insert() and heap_update() optimization