Date: Wed, 31 Aug 2011 12:11:02 +0400 From: Lev Serebryakov <lev@FreeBSD.org> To: Kirk McKusick <mckusick@mckusick.com> Cc: freebsd-fs@FreeBSD.org, lev@FreeBSD.org Subject: Re: Very inconsistent (read) speed on UFS2 Message-ID: <170569583.20110831121102@serebryakov.spb.ru> In-Reply-To: <201108302300.p7UN0jJ6022811@chez.mckusick.com> References: <1693072185.20110831022933@serebryakov.spb.ru> <201108302300.p7UN0jJ6022811@chez.mckusick.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hello, Kirk. You wrote 31 =E0=E2=E3=F3=F1=F2=E0 2011 =E3., 3:00:45: > So, my new theory on why your O_DIRECT test is running slowly is due > to the single threading in the GEOM layer. Pawel Jakub Dawidek (pjd@) > gave a very interesting talk on this problem at this year's BSDCan. I want to stress my point: it is not low speed, what especially bother me. It is inconsistency. Bad drive, always-single-threaded-GEOM, etc., should give consistent slow speed. Bad code in geom_raid5 could give inconsistent write speed, du to caching, but reading path is as straight and simple as possible here. Bad drive or AST-GEOM is not what simple to fix (ok, bad drive IS simple to replace, of course). But I suspect, that here is some simple "misunderstanding" between geom_raid5 code and VFS/FFS2 layer, which I could fix. But I don't know where to look at. --=20 // Black Lion AKA Lev Serebryakov <lev@FreeBSD.org>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?170569583.20110831121102>