Date: Mon, 28 Jan 2013 16:12:28 +0000 From: Laurence Gill <laurencesgill@googlemail.com> To: freebsd-fs@freebsd.org Subject: Re: HAST performance overheads? Message-ID: <20130128161228.477ce174@googlemail.com> In-Reply-To: <20130128120055.6ca7c734@googlemail.com> References: <20130125121044.1afac72e@googlemail.com> <20130127134845.GC1346@garage.freebsd.pl> <20130128120055.6ca7c734@googlemail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Mon, 28 Jan 2013 12:00:55 +0000 Laurence Gill <laurencesgill@googlemail.com> wrote: > On Sun, 27 Jan 2013 14:48:46 +0100 > Pawel Jakub Dawidek <pjd@FreeBSD.org> wrote: > > > > 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=/dev/zero of=/dev/da0 bs=1m count=10240 > > > > then configure HAST and: > > > > # dd if=/dev/zero of=/dev/hast/disk0 bs=1m count=10240 > > > > 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=16k count=65535' case. Let try raw providers first. > > > > Thanks for the reply. I'm using FreeBSD 9.1-RELEASE. Here are the > results: > > # dd if=/dev/zero of=/dev/da0 bs=1m count=10240 > 10737418240 bytes transferred in 755.144644 secs (14219022 bytes/sec) > > # dd if=/dev/zero of=/dev/hast/disk0 bs=1m count=10240 > 10737418240 bytes transferred in 844.167602 secs (12719534 bytes/sec) > > > Which indicates a very small overhead, hmmm... > Further to this, sticking with the 1 disk for testing, I see the following: - UFS on da0 # dd if=/dev/zero of=test.dat bs=1m count=10240 10737418240 bytes transferred in 76.112873 secs (141072302 bytes/sec) - UFS on hast/disk0 # dd if=/dev/zero of=test.dat bs=1m count=10240 10737418240 bytes transferred in 855.720985 secs (12547803 bytes/sec) Which is roughly the same as using the raw hast provider. - zfs on da0 # dd if=/dev/zero of=test.dat bs=1m count=10240 10737418240 bytes transferred in 114.338900 secs (93908707 bytes/sec) - zfs on hast/disk0 # dd if=/dev/zero of=test.dat bs=1m count=10240 10737418240 bytes transferred in 1287.088416 secs (8342409 bytes/sec) Which seems slower than the raw provider by approx 4MB/s. So I'm still trying to work out why the extra "drop" when using ZFS on hast... - -- Laurence Gill f: 08721 157 665 skype: laurencegg e: laurencesgill@googlemail.com PGP on Key Servers -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.19 (GNU/Linux) iEYEARECAAYFAlEGo3QACgkQygVt8Sq0Pf8K3QCfVA+nofIgRHM/gYiAzis6TF5+ VvYAn2kEOVtGySR0eZtegGrvUap5BVhx =9fCv -----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130128161228.477ce174>