Re: Different compression methods for FPI

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Andrey Borodin <x4mmm(at)yandex-team(dot)ru>
Cc: Justin Pryzby <pryzby(at)telsasoft(dot)com>, Peter Eisentraut <peter(dot)eisentraut(at)enterprisedb(dot)com>, Heikki Linnakangas <hlinnaka(at)iki(dot)fi>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>, Thomas Munro <thomas(dot)munro(at)gmail(dot)com>, Peter Geoghegan <pg(at)bowt(dot)ie>, Andres Freund <andres(at)anarazel(dot)de>
Subject: Re: Different compression methods for FPI
Date: 2021-06-17 06:44:26
Message-ID: YMrvSry/
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

On Thu, Jun 17, 2021 at 10:19:47AM +0900, Michael Paquier wrote:
> Yeah, I am tempted to just add LZ4 as a first step as the patch
> footprint would be minimal, and we could come back to zstd once we
> have more feedback from the field, if that's necessary. As said
> upthread, we have more flexibility with WAL than for the relation
> data.

I have worked more on that today and finished with two patches:
- 0001 is the mininal patch to add support for LZ4. This is in a
rather committable shape. I noticed that we checked for an incorrect
error code in the compression and decompression paths as LZ4 APIs can
return a negative result. There were also some extra bugs I spotted.
Its size is satisfying for what it does, and there is MSVC support
12 files changed, 176 insertions(+), 48 deletions(-)
- 0002 is the extra code need to add ZSTD and do the same. This still
requires support for MSVC and I have not checked the internals of ZSTD
to see if we do the compress/decompress calls the right way.

While on it, I am going to switch my buildfarm animal to use LZ4 for
toast.. Just saying.

In response to


Browse pgsql-hackers by date

  From Date Subject
Next Message Andrey Borodin 2021-06-17 06:45:37 Re: Different compression methods for FPI
Previous Message Amit Kapila 2021-06-17 06:41:53 Re: Decoding speculative insert with toast leaks memory