Re: Different compression methods for FPI - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: Different compression methods for FPI
Date
Msg-id YMryVgORHbhRGSDp@paquier.xyz
Whole thread Raw
In response to Re: Different compression methods for FPI  (Andrey Borodin <x4mmm@yandex-team.ru>)
Responses Re: Different compression methods for FPI
List pgsql-hackers
On Thu, Jun 17, 2021 at 11:45:37AM +0500, Andrey Borodin wrote:
> Konstantin, Daniil and Justin are working on compressing libpq
> [0]. That would make walsender compress WAL automatically.
> And we (at least I and Dan) are inclined to work on compressing
> on-disk WAL as soon as libpq compression will be committed.

What's the relationship between libpq and WAL?  Just the addition of
zstd in the existing dependency chain?

> Zstd is much better at compressing long data sequences than lz4.

That's my impression.

> I'm sure we will need such codec in core one day.
>
> Because compressing sequence of zeroes is cheap, even for pglz. But
> we still need to support 'no compression at all', this mode benefits
> from hole-in-a-page a lot. Until we send and store WAL uncompressed,
> of cause.

I am not sure, but surely this will come up in future discussions as a
separate problem.
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Andrey Borodin
Date:
Subject: Re: Different compression methods for FPI
Next
From: Fabien COELHO
Date:
Subject: Re: pgbench logging broken by time logic changes