Re: Strange IO performance with UFS
--ukxn79E1+tniTzNx
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
On Sun, Jul 06, 2014 at 01:13:29AM -0700, Adrian Chadd wrote:
> On 6 July 2014 00:17, Konstantin Belousov <kostikbel@gmail.com> wrote:
> > On Sat, Jul 05, 2014 at 05:34:48PM -0700, Adrian Chadd wrote:
> >> Hm, wait a sec. So if the IO size is a multiple of the underlying FS
> >> block size, it should be okay?
> > It was already suggested in the thread, and confirmed later.
> > Did you read the mails ?
>=20
> I did, I must've missed it. SOrry!
The thread was definitely long.
I am slightly curious how other systems perform in this case.
In particular, since fio is originating from Linux, it probably
should demonstrate some good behaviour with the default settings.
--ukxn79E1+tniTzNx
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2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=JtJc
-----END PGP SIGNATURE-----
--ukxn79E1+tniTzNx--
討論串 (同標題文章)
完整討論串 (本文為第 17 之 22 篇)：