Re: Two fsync related performance issues? - Mailing list pgsql-hackers

From Paul Guo
Subject Re: Two fsync related performance issues?
Date
Msg-id CAEET0ZERL+UGw9aG6zh=yF7tcnVB5g1UbFE8+v9DvVwhwTgo9w@mail.gmail.com
Whole thread Raw
In response to Re: Two fsync related performance issues?  (Michael Paquier <michael@paquier.xyz>)
Responses Re: Two fsync related performance issues?
List pgsql-hackers
Thanks for the replies.

On Tue, May 12, 2020 at 2:04 PM Michael Paquier <michael@paquier.xyz> wrote:
On Tue, May 12, 2020 at 12:55:37PM +0900, Fujii Masao wrote:
> On 2020/05/12 9:42, Paul Guo wrote:
>> 1. StartupXLOG() does fsync on the whole data directory early in
>> the crash recovery. I'm wondering if we could skip some
>> directories (at least the pg_log/, table directories) since wal,
>> etc could ensure consistency.
>
> I agree that we can skip log directory but I'm not sure if skipping
> table directory is really safe. Also ISTM that we can skip the directories
> that those contents are removed or zeroed during recovery,
> for example, pg_snapshots, pg_substrans, etc.

Basically excludeDirContents[] as of basebackup.c.

table directories & wal fsync probably dominates the fsync time. Do we
know any possible real scenario that requires table directory fsync?


pgsql-hackers by date:

Previous
From: Daniel Gustafsson
Date:
Subject: Vintage unused variables in pg_dump.c
Next
From: Pavel Stehule
Date:
Subject: proposal - plpgsql - FOR over unbound cursor