Date: Sun, 27 Jan 2013 14:48:46 +0100 From: Pawel Jakub Dawidek <pjd@FreeBSD.org> To: Laurence Gill <laurencesgill@googlemail.com> Cc: freebsd-fs@freebsd.org Subject: Re: HAST performance overheads? Message-ID: <20130127134845.GC1346@garage.freebsd.pl> In-Reply-To: <20130125121044.1afac72e@googlemail.com> References: <20130125121044.1afac72e@googlemail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--WplhKdTI2c8ulnbP
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
On Fri, Jan 25, 2013 at 12:10:44PM +0000, Laurence Gill wrote:
> If I create ZFS raidz2 on these...
>=20
> - # zpool create pool raidz2 da0 da1 da2 da3 da4 da5
>=20
> Then run a dd test, a sample output is...
>=20
> - # dd if=3D/dev/zero of=3Dtest.dat bs=3D1M count=3D1024
> 1073741824 bytes transferred in 7.689634 secs (139634974 bytes/sec)
>=20
> - # dd if=3D/dev/zero of=3Dtest.dat bs=3D16k count=3D65535
> 1073725440 bytes transferred in 1.909157 secs (562408130 bytes/sec)
>=20
> This is much faster than compared to running hast, I would expect an
> overhead, but not this much. For example:
>=20
> - # hastctl create disk0/disk1/disk2/disk3/disk4/disk5
> - # hastctl role primary all
> - # zpool create pool raidz2 disk0 disk1 disk2 disk3 disk4 disk5
>=20
> Run a dd test, and the speed is...
>=20
> - # dd if=3D/dev/zero of=3Dtest.dat bs=3D1M count=3D1024
> 1073741824 bytes transferred in 40.908153 secs (26247624 bytes/sec)
>=20
> - # dd if=3D/dev/zero of=3Dtest.dat bs=3D16k count=3D65535
> 1073725440 bytes transferred in 42.017997 secs (25553942 bytes/sec)
Let's try to test one step at a time. Can you try to compare sequential
performance of regular disk vs. HAST with no secondary configured?
By no secondary configured I mean 'remote' set to 'none'.
Just do:
# dd if=3D/dev/zero of=3D/dev/da0 bs=3D1m count=3D10240
then configure HAST and:
# dd if=3D/dev/zero of=3D/dev/hast/disk0 bs=3D1m count=3D10240
Which FreeBSD version is it?
PS. Your ZFS tests are pretty meaningless, because it is possible that
everything will end up in memory. I'm sure this is what happens in
'bs=3D16k count=3D65535' case. Let try raw providers first.
--=20
Pawel Jakub Dawidek http://www.wheelsystems.com
FreeBSD committer http://www.FreeBSD.org
Am I Evil? Yes, I Am! http://tupytaj.pl
--WplhKdTI2c8ulnbP
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (FreeBSD)
iEYEARECAAYFAlEFMD0ACgkQForvXbEpPzRMigCglS8ZP9RggVl0MfVk+A25xgd2
29wAnigH5gA4RXxKI/4XLfKT8sW9eoPP
=D2zj
-----END PGP SIGNATURE-----
--WplhKdTI2c8ulnbP--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130127134845.GC1346>