Horton

Re: Teach pg_receivewal to use lz4 compression

gkokolatos@pm.me

View thread View on postgr.es
From: gkokolatos@pm.me
To: Michael Paquier <michael@paquier.xyz>
Cc: Robert Haas <robertmhaas@gmail.com>, Magnus Hagander <magnus@hagander.net>, Dilip Kumar <dilipbalaut@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2021-11-03T08:05:53Z
Lists: pgsql-hackers 

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐

On Wednesday, November 3rd, 2021 at 12:23 AM, Michael Paquier <michael@paquier.xyz> wrote:

> On Tue, Nov 02, 2021 at 12:31:47PM -0400, Robert Haas wrote:
>> On Tue, Nov 2, 2021 at 8:17 AM Magnus Hagander magnus@hagander.net wrote:
>>> I think for the end user, it is strictly better to name it "gzip",
>>> and given that the target of this option is the end user we should
>>> do so. (It'd be different it we were talking about a build-time
>>> parameter to configure).
>>
>> I agree. Also, I think there's actually a file format called "zlib"
>> which is slightly different from the "gzip" format, and you have to be
>> careful not to generate the wrong one.
>
> Okay, fine by me.  It would be better to be also consistent in
> WalCompressionMethods once we switch to this option value, then.

I will revert to gzip for version 9. Should be out shortly.

Cheers,
//Georgios
>
> Michael

Commits

  1. Fix header inclusion order in pg_receivewal.c

  2. Remove useless LZ4 system call on failure when writing file header

  3. Add support for LZ4 compression in pg_receivewal

  4. Rework compression options of pg_receivewal

  5. Clarify some errors in pg_receivewal when closing WAL segments