Date: Thu, 24 Mar 2016 08:16:45 -0400 From: Paul Mather <paul@gromit.dlib.vt.edu> To: Warner Losh <imp@bsdimp.com> Cc: Oliver Psotta <oliver.psotta@posteo.de>, freebsd-arm@freebsd.org Subject: Re: Effect of partitioning on wear-leveling Message-ID: <D7533887-E13D-4134-A982-F8158FA96CE0@gromit.dlib.vt.edu> In-Reply-To: <CANCZdfoefzvP5g5hJjiRnUuzB1K_bhKEHeyMJU-V2B0MyeAUgA@mail.gmail.com> References: <20160321175952.GA83908@www.zefox.net> <1458586884.68920.96.camel@freebsd.org> <20160321221153.GB83908@www.zefox.net> <1458600070.68920.107.camel@freebsd.org> <1973487B-0AA7-468D-A9CC-319FBE2122F0@netgate.com> <CANCZdfrCWXAswe02Qd3tTiDL8O_4TGEWbhFqgft4Q9aKj7ixvg@mail.gmail.com> <20160322033417.GD83908@www.zefox.net> <201603230349.VAA20311@mail.lariat.net> <CANCZdfp5jffpHcnoDJg24stUydEssASeC4owmz7n-fmY=evGzQ@mail.gmail.com> <AC03ED3F-D113-4640-9DFD-DCAC193A5517@posteo.de> <CANCZdfoefzvP5g5hJjiRnUuzB1K_bhKEHeyMJU-V2B0MyeAUgA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mar 24, 2016, at 12:05 AM, Warner Losh <imp@bsdimp.com> wrote: > When you have a fleet of thousands of ssds, you'll get failures no = matter > the quality... Pertinent to discussion of SSD failures is this article about the topic, = which summarises a FAST 2016 paper on the subject: = http://www.zdnet.com/article/ssd-reliability-in-the-real-world-googles-exp= erience/ Here are the "key conclusions" from the ZDNet article (and I quote): "=E2=80=A2 Ignore Uncorrectable Bit Error Rate (UBER) specs. A = meaningless number. =E2=80=A2 Good news: Raw Bit Error Rate (RBER) increases slower than = expected from wearout and is not correlated with UBER or other failures. =E2=80=A2 High-end SLC drives are no more reliable that MLC drives. =E2=80=A2 Bad news: SSDs fail at a lower rate than disks, but UBER rate = is higher (see below for what this means). =E2=80=A2 SSD age, not usage, affects reliability. =E2=80=A2 Bad blocks in new SSDs are common, and drives with a large = number of bad blocks are much more likely to lose hundreds of other = blocks, most likely due to die or chip failure. =E2=80=A2 30-80 percent of SSDs develop at least one bad block and 2-7 = percent develop at least one bad chip in the first four years of = deployment." Cheers, Paul. >=20 > Warner > On Mar 23, 2016 1:55 AM, "Oliver Psotta" <oliver.psotta@posteo.de> = wrote: >=20 >> Which SSDs failed on you, Warner? There sure are some rotten apples, >> but the Samsung 840 pro, for example, were (are) quite reliable. >>=20 >> -Oliver >>=20 >>> On 23 Mar 2016, at 07:45, Warner Losh <imp@bsdimp.com> wrote: >>>=20 >>> Hope your SSDs are better at reporting things than ours. We've seen = some >>> SSDs >>> just fail even though the previous SMART data said we've used maybe = 20% >> of >>> the >>> drive's write ability.... >>=20 >> _______________________________________________ >> freebsd-arm@freebsd.org mailing list >> https://lists.freebsd.org/mailman/listinfo/freebsd-arm >> To unsubscribe, send any mail to = "freebsd-arm-unsubscribe@freebsd.org" >>=20 > _______________________________________________ > freebsd-arm@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-arm > To unsubscribe, send any mail to "freebsd-arm-unsubscribe@freebsd.org" >=20
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?D7533887-E13D-4134-A982-F8158FA96CE0>