Subject: Re: Questionaire: Common WAL write rates on busy servers.
To: Andres Freund <andres@anarazel.de>
References: <20170425041743.ddabv2mwevcmtx5a@alap3.anarazel.de>
 <20170427155906.uudovtd62x56brmx@alap3.anarazel.de>
 <a842fbdf-b761-bf21-47dd-cd87d05a8d04@commandprompt.com>
 <20170427163407.dtqi6v57bj4tiu5c@alap3.anarazel.de>
 <80f280c1-910f-7eae-33a2-fd0cfc9ee8c9@commandprompt.com>
 <20170427173540.fqaucmclpowfy5n7@alap3.anarazel.de>
 <cc3c764c-26a7-315c-5d81-7a84e380277a@2ndquadrant.com>
 <20170427233433.eg5vwlvbqyc2nruh@alap3.anarazel.de>
Cc: "Joshua D. Drake" <jd@commandprompt.com>, pgsql-general@postgresql.org,
 pgsql-performance@postgresql.org
From: Tomas Vondra <tomas.vondra@2ndquadrant.com>
Message-ID: <48afdba3-3429-de92-6fa8-36e2ebe5fd03@2ndquadrant.com>
Date: Sat, 29 Apr 2017 02:41:19 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101
 Thunderbird/45.4.0
MIME-Version: 1.0
In-Reply-To: <20170427233433.eg5vwlvbqyc2nruh@alap3.anarazel.de>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk
Sender: pgsql-performance-owner@postgresql.org

On 04/28/2017 01:34 AM, Andres Freund wrote:
> On 2017-04-28 01:29:14 +0200, Tomas Vondra wrote:
>> I can confirm this observation. I bought the Intel 750 NVMe SSD last year,
>> the device has 1GB DDR3 cache on it (power-loss protected), can do ~1GB/s of
>> sustained O_DIRECT sequential writes. But when running pgbench, I can't push
>> more than ~300MB/s of WAL to it, no matter what I do because of
>> WALWriteLock.
>
> Hm, interesting.  Even if you up wal_buffers to 128MB, use
> synchronous_commit = off, and play with wal_writer_delay/flush_after?
>

I think I've tried things like that, but let me do some proper testing. 
I'll report the numbers in a few days.

regards

-- 
Tomas Vondra                  http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services


-- 
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance