Date: Wed, 27 Oct 2010 04:19:04 -0700 From: David Wolfskill <david@catwhisker.org> To: freebsd-performance@freebsd.org Subject: Re: Possible evidence of performance regression for 8.1-S (vs. 7.1) Message-ID: <20101027111904.GF9443@albert.catwhisker.org> In-Reply-To: <ia9137$gag$1@dough.gmane.org> References: <20101026121352.GC2262@albert.catwhisker.org> <AANLkTi=Q9EwPJ%2BC4n3UFF9uFtNsukntkmqSjrXsHE3cm@mail.gmail.com> <20101020174854.GZ21226@albert.catwhisker.org> <20101021215330.GA86224@dan.emsphone.com> <20101021224237.GG52404@albert.catwhisker.org> <4CC6C396.1010905@freebsd.org> <20101026174501.GH2262@albert.catwhisker.org> <ia8srt$th0$1@dough.gmane.org> <20101027105506.GD9443@albert.catwhisker.org> <ia9137$gag$1@dough.gmane.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--kR3zbvD4cgoYnS/6 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Wed, Oct 27, 2010 at 01:06:18PM +0200, Ivan Voras wrote: > On 10/27/10 12:55, David Wolfskill wrote: >=20 > > That *is* a problem, as I cannot justify a migration to a branch > > of FreeBSD that imposes about a 23% penalty in elapsed time on this > > workload. I want folks at work to have more reason to want to use > > (newer branches of) FreeBSD, not less. >=20 > That is understandable, :-} > I'm only saying that given the symptoms - some > "waiting" that is not accounted in system or user time and with NFS > mostly ruled out (maybe a better test would be to extract a local > tarball *to* the NFS server), I could run that test, sure.... > the next suspect is the disk system. In > case you never tried "diskinfo -vt", its results look like this: >=20 > outside: 102400 kbytes in 1.222998 sec =3D 83729 kbytes/sec > middle: 102400 kbytes in 1.448580 sec =3D 70690 kbytes/sec > inside: 102400 kbytes in 2.531694 sec =3D 40447 kbytes/sec >=20 > note 2x drop in performance between outer and inner tracks. OK, but I'm not sure how that's likely to work for a multi-spindle RAID 0 group.... > It's not > certain, especially since your results are repeatable and you always use > exactly the same file system, For the NFS tests just done. For the "real" tests, the same file system was used when I ran the tests on the same machine; (obviously) different machines have different (local) file systems. (Still, similarly- configured 4-spindle RAID 0 groups were used for the target file systems on each machine.) > but it's possible. Have you tried the sysctls I've posted? Err... Not yet, no. I just got up & read your message. I leave for work in just over 3 hours. :-} [While it is possible to login from home, it's a liitle more awkward than (say) logging in to the FreeBSD cluster, and this is the time of day when I reboot my laptop a few times -- it's presently running 9.0-CURRENT #23 r214372 & building 9.0-CURRENT #23 r214413; it just finished building & smoke-testing FreeBSD 8.1-STABLE #14 r214413.] Peace, david --=20 David H. Wolfskill david@catwhisker.org Depriving a girl or boy of an opportunity for education is evil. See http://www.catwhisker.org/~david/publickey.gpg for my public key. --kR3zbvD4cgoYnS/6 Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (FreeBSD) iEYEARECAAYFAkzICqgACgkQmprOCmdXAD0fXACfWVaJ/a66ZkX0dsQYpw38nqr4 nBAAnR59+aF3mBCGJ+Iz3S0mun42FFIn =U0fT -----END PGP SIGNATURE----- --kR3zbvD4cgoYnS/6--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20101027111904.GF9443>