Date: Sun, 11 Mar 2001 03:11:42 +1100 (EST) From: Bruce Evans <bde@zeta.org.au> To: Warner Losh <imp@harmony.village.org> Cc: Josef Karthauser <joe@tao.org.uk>, "David O'Brien" <obrien@FreeBSD.org>, bde@FreeBSD.org, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/release/scripts doFS.sh Message-ID: <Pine.BSF.4.21.0103110311120.14887-100000@besplex.bde.org> In-Reply-To: <200103091915.f29JFdI10005@harmony.village.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 9 Mar 2001, Warner Losh wrote: > In message <20010309182154.B761@tao.org.uk> Josef Karthauser writes: > > Trust me, it made a difference. :-) > > In fact I stuck in a `tunef -p' when I was testing this and found w/o > > -m X, X < 6; would still show the default of 8%. Someone mentioned that > > there may be a bug where `tunefs' will not set the value to 0. > > tunefs will not do -m 0. It makes sure that minfree is at least 1 > before setting it. Bug or feature? Who knows. FUD. `tunefs -m 0' works perfectly. The problem seems to be that people ignore the warning about setting minfree to <= 5 without also setting optimization to `space'. ffs changes the optimization dynamically but doesn't (and can't reasonably) do it perfectly. Allocating whole blocks for small files while the optimization is set to `time' gives lots of fragments. Then dynamically changing the setting to `space' only helps much if there are plenty of small files to fill the fragments. Any remaining whole blocks would soon be used up for large files if the disk gets nearly full. All the small files would have to be repacked to give the same result as starting with space optimization, but this would take too long. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0103110311120.14887-100000>