Re: GUC-ify walsender MAX_SEND_SIZE constant

From: Andres Freund <andres(at)anarazel(dot)de>
To: Majid Garoosi <amoomajid99(at)gmail(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: GUC-ify walsender MAX_SEND_SIZE constant
Date: 2024-02-09 19:03:46
Message-ID: 20240209190346.22fwj4xb67cmh3io@awork3.anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

On 2024-01-19 23:04:50 +0330, Majid Garoosi wrote:
> Following is the description which is also written in the commit message:
> MAX_SEND_SIZE parameter was used in WALSender to limit maximum size of
> a WAL data packet sent to a WALReceiver during replication. Although
> its default value (128kB) was a reasonable value, it was written in
> 2010. Since the hardwares have changed a lot since then, a PostgreSQL
> user might need to customize this value.
> For example, if a database's disk has a high bandwidth and a high
> latency at the same time, it makes more sense to read bigger chunks of
> data from disk in each iteration. One example of such disks is a remote
> disk. (e.g. an RBD volume)

The way we read the WAL data is perfectly prefetchable by the the OS, so I
wouldn't really expect gains here. Have you actually been able to see a
performance benefit by increasing MAX_SEND_SIZE?

I don't think we should add configuration parameters without at least some
demonstration of practical gains, otherwise we'll end up with hundreds of
never-useful config options.

Greetings,

Andres Freund

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Dave Cramer 2024-02-09 19:23:46 Re: [PATCH] Add native windows on arm64 support
Previous Message Andrey Borodin 2024-02-09 19:02:08 Re: glibc qsort() vulnerability