Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 18 May 2016 08:49:34 -0500 (CDT)
From:      Bob Friesenhahn <bfriesen@simple.dallas.tx.us>
To:        Alex Tutubalin <lexa@lexa.ru>
Cc:        "freebsd-fs@freebsd.org" <freebsd-fs@freebsd.org>
Subject:   Re: ZFS performance bottlenecks: CPU or RAM or anything else?
Message-ID:  <alpine.GSO.2.20.1605180843280.7756@freddy.simplesystems.org>
In-Reply-To: <f1cc3ee5-c141-b366-83bf-3ee0179381bf@lexa.ru>
References:  <8441f4c0-f8d1-f540-b928-7ae60998ba8e@lexa.ru> <f87ec54a-104e-e712-7793-86c37285fdaa@internetx.com> <16e474da-6b20-2e51-9981-3c262eaff350@lexa.ru> <BD7DE274-04EB-4B19-988D-5A6FADC5B51A@digsys.bg> <1e012e43-a49b-6923-3f0a-ee77a5c8fa70@lexa.ru> <86shxgsdzh.fsf@WorkBox.Home> <CAHEMsqZto0wD9Ko4E9YUpYvea4jM0E4f2nC1HkAwcCG=6DfX-A@mail.gmail.com> <f1cc3ee5-c141-b366-83bf-3ee0179381bf@lexa.ru>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 18 May 2016, Alex Tutubalin wrote:
>
> If so, raidz will have huge write performance benefit in my case: single 
> write of one large file.

This is not proven in practice.  With mirrors one typically has more 
vdevs and each vdev gets a zfs block-size write in turn, using a 
round robin agorithm (tuned for available space vdev).  Drive IOPs are 
saved since the blocks are not diced into smaller fragments (as raidzN 
requires).  With raidz it is necessary to also pay the cost of the 
parity computations, which are not needed with mirroring.

Bob
-- 
Bob Friesenhahn
bfriesen@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?alpine.GSO.2.20.1605180843280.7756>