Date: Mon, 19 Jan 2004 17:49:15 -0700 (MST) From: Warren Block <wblock@wonkity.com> To: "Cordula's Web" <cpghost@cordula.ws> Cc: freebsd-questions@freebsd.org Subject: Re: [was: Cutting the power ... ] Journaling file system Message-ID: <20040119172643.S84590@wonkity.com> In-Reply-To: <20040119074947.28D7440854@fw.farid-hajji.net> References: <007701c3ddaf$8c58f000$8b01010a@nyalaptopen> <1074484137.21654.7.camel@tux> <20040119074947.28D7440854@fw.farid-hajji.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 19 Jan 2004, Cordula's Web wrote: > Beware when using flash ram as some kind of live filesystem! > Flash media has a quite limited number of erase cycles > You may think that's plenty, but since filesystem meta-data > is often written at the same location (superblocks etc...), > this location will quickly reach the threshold, where it > can't be used anymore! Current flash memory spreads the writes around through "wear leveling." A logical write to an address is remapped internally to keep the wear as even as possible. There's a kind of neat calculator at: http://www.m-sys.com/content/Developer/Calc.asp There's also a PDF from the same company that talks about how it works, but I couldn't find it on their web site, so: http://www.spezial.de/commercio/dateien/magazin/FFD_Life_Expectancy.pdf -Warren Block * Rapid City, South Dakota USA
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040119172643.S84590>