Date: Fri, 22 Aug 2014 14:24:51 -0400 From: Paul Kraus <paul@kraus-haus.org> To: Daniel Feenberg <feenberg@nber.org> Cc: RW <rwmaillists@googlemail.com>, freebsd-questions@freebsd.org Subject: Re: solid state drives? Message-ID: <7BDE9B34-C73C-4B29-A9BD-53228336BE70@kraus-haus.org> In-Reply-To: <alpine.LRH.2.11.1408221239310.9489@nber2.nber.org> References: <53F22E89.3050005@rcn.com> <53F2399D.5050609@hiwaay.net> <20140822170112.69830ad9@gumby.homeunix.com> <alpine.LRH.2.11.1408221239310.9489@nber2.nber.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Aug 22, 2014, at 12:41, Daniel Feenberg <feenberg@nber.org> wrote: > I sort of understand that - but does the SSD have the ability to move = unchanged data around to even out the wear? That is, if I fill the drive = with 100GB of never changing files, and then write lots of frequently = changing files to the last 20GB, does this put all the wear on a small = portion of the drive, while most of the drive suffers no wear at all? = Maybe I should do a full backup and restore once a year? Keep in mind that location is not a physical parameter in an SSD. Better = (all today ?) SSD=92s do wear leveling where writes are committed to the = cells that have the lowest write counts. Remember, writes count towards = wear out while I do not think reads do. So an SSD that has write once = data (and archive), should never wear out. So it does not matter where within the block range you write, the SSD = puts it where it wants :-) -- Paul Kraus paul@kraus-haus.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?7BDE9B34-C73C-4B29-A9BD-53228336BE70>