Date: Tue, 04 Nov 2014 15:31:12 -0800 From: Darren Pilgrim <list_freebsd@bluerosetech.com> To: Chris H <bsd-lists@bsdforge.com> Cc: freebsd-ports@freebsd.org Subject: Re: Reducing the size of the ports tree (brainstorm v2) Message-ID: <545961C0.9070303@bluerosetech.com> In-Reply-To: <238b9bc5e8987fe4fa2e9de3f68dc2ac@ultimatedns.net> References: <20141031185621.GC15967@ivaldir.etoilebsd.net> <54573B31.7080809@gmx.de>, <20141103212438.0893c3dc@kalimero.tijl.coosemans.org>, <14d0c0b9ee9ca31877d43a3c29481717@ultimatedns.net> <238b9bc5e8987fe4fa2e9de3f68dc2ac@ultimatedns.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On 11/4/2014 2:28 PM, Chris H wrote: > On Tue, 04 Nov 2014 13:21:31 -0800 "Chris H" <bsd-lists@bsdforge.com> wrote >> gpart(8) -a gives you what you need. If it's truly as bad as all that, >> mounting the ports tree on a 512k aligned slice will reduce the "slack" > ahem... > that was s/512k/512b/g The issue of 512b sector storage media going underlies this discussion. 4k drives are the new typical. Flash uses even larger block sizes. Using an alignment of less than one sector yields significant performance penalties when doing small reads or writes. The on-disk size of the ports tree more than doubles when using 4k blocks because all those files use 4k to store what often fits in 512b. Cutting down the number of files has wide-reaching performance gains with subversion as well.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?545961C0.9070303>