Skip site navigation (1) Skip section navigation (2)

Re: Inserting 8MB bytea: just 25% of disk perf used?

From: Ivan Voras <ivoras(at)freebsd(dot)org>
To: pgsql-performance(at)postgresql(dot)org
Subject: Re: Inserting 8MB bytea: just 25% of disk perf used?
Date: 2010-01-19 11:52:24
Message-ID: hj46di$9bi$1@ger.gmane.org (view raw or flat)
Thread:
Lists: pgsql-performance
On 01/19/10 11:16, fkater(at)googlemail(dot)com wrote:
> fkater(at)googlemail(dot)com:
>
>> I'll try to execute these tests on a SSD
>> and/or Raid system.
>
> FYI:
>
> On a sata raid-0 (mid range hardware) and recent 2x 1.5 TB
> disks with a write performance of 100 MB/s (worst, to 200
> MB/s max), I get a performance of 18.2 MB/s. Before, with
> other disk 43 MB/s (worst to 70 MB/s max) postgres came to
> 14-16 MB/s.

[I just skimmed this thread - did you increase the number of WAL logs to 
something very large, like 128?]

> So, I conclude finally:
>
> (1) Postgresql write throughput (slowly) scales with the
> harddisk speed.
>
> (2) The throughput (not counting WAL doubling data) in
> postgresql is 20-25% of the disk thoughput.

And this is one of the more often forgot reasons why storing large 
objects in a database rather than in the file systems is a bad idea :)



In response to

Responses

pgsql-performance by date

Next:From: Richard HuxtonDate: 2010-01-19 12:16:54
Subject: Re: Inserting 8MB bytea: just 25% of disk perf used?
Previous:From: fkater@googlemail.comDate: 2010-01-19 10:55:26
Subject: Re: Inserting 8MB bytea: just 25% of disk perf used?

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group