From: Phillip Susi <psusi@cfl.rr.com>
To: Michael Loftis <mloftis@wgops.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: FYI: RAID5 unusably unstable through 2.6.14
Date: Tue, 17 Jan 2006 19:21:45 -0500 [thread overview]
Message-ID: <43CD8A19.3010100@cfl.rr.com> (raw)
In-Reply-To: <B34375EBA93D2866BECF5995@d216-220-25-20.dynip.modwest.com>
Your understanding of statistics leaves something to be desired. As you
add disks the probability of a single failure is grows linearly, but the
probability of double failure grows much more slowly. For example:
If 1 disk has a 1/1000 chance of failure, then
2 disks have a (1/1000)^2 chance of double failure, and
3 disks have a (1/1000)^2 * 3 chance of double failure
4 disks have a (1/1000)^2 * 7 chance of double failure
Thus the probability of double failure on this 4 drive array is ~142
times less than the odds of a single drive failing. As the probably of
a single drive failing becomes more remote, then the ratio of that
probability to the probability of double fault in the array grows
exponentially.
( I think I did that right in my head... will check on a real calculator
later )
This is why raid-5 was created: because the array has a much lower
probabiliy of double failure, and thus, data loss, than a single drive.
Then of course, if you are really paranoid, you can go with raid-6 ;)
Michael Loftis wrote:
> Absolutely not. The more spindles the more chances of a double failure.
> Simple statistics will mean that unless you have mirrors the more drives
> you add the more chance of two of them (really) failing at once and
> choking the whole system.
>
> That said, there very well could be (are?) cases where md needs to do a
> better job of handling the world unravelling.
> -
next prev parent reply other threads:[~2006-01-18 0:21 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-01-17 19:35 FYI: RAID5 unusably unstable through 2.6.14 Cynbe ru Taren
2006-01-17 19:39 ` Benjamin LaHaise
2006-01-17 20:13 ` Martin Drab
2006-01-17 23:39 ` Michael Loftis
2006-01-18 2:30 ` Martin Drab
2006-02-02 20:33 ` Bill Davidsen
2006-02-03 0:57 ` Martin Drab
2006-02-03 1:13 ` Martin Drab
2006-02-03 15:41 ` Phillip Susi
2006-02-03 16:13 ` Martin Drab
2006-02-03 16:38 ` Phillip Susi
2006-02-03 17:22 ` Roger Heflin
2006-02-03 19:38 ` Phillip Susi
2006-02-03 17:51 ` Martin Drab
2006-02-03 19:10 ` Roger Heflin
2006-02-03 19:12 ` Martin Drab
2006-02-03 19:41 ` Phillip Susi
2006-02-03 19:45 ` Martin Drab
2006-01-17 19:56 ` Kyle Moffett
2006-01-17 19:58 ` David R
2006-01-17 20:00 ` Kyle Moffett
2006-01-17 23:27 ` Michael Loftis
2006-01-18 0:12 ` Kyle Moffett
2006-01-18 11:24 ` Erik Mouw
2006-01-18 0:21 ` Phillip Susi [this message]
2006-01-18 0:29 ` Michael Loftis
2006-01-18 2:10 ` Phillip Susi
2006-01-18 3:01 ` Michael Loftis
2006-01-18 16:49 ` Krzysztof Halasa
2006-01-18 16:47 ` Krzysztof Halasa
2006-02-02 22:10 ` Bill Davidsen
2006-02-08 21:58 ` Pavel Machek
2006-01-18 10:54 ` Helge Hafting
2006-01-18 16:15 ` Mark Lord
2006-01-18 17:32 ` Alan Cox
2006-01-19 15:59 ` Mark Lord
2006-01-19 16:25 ` Alan Cox
2006-02-08 14:46 ` Alan Cox
2006-01-18 23:37 ` Neil Brown
2006-01-19 15:53 ` Mark Lord
2006-01-19 0:13 ` Neil Brown
2006-02-03 17:00 Salyzyn, Mark
2006-02-03 17:39 ` Martin Drab
2006-02-03 19:46 ` Phillip Susi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=43CD8A19.3010100@cfl.rr.com \
--to=psusi@cfl.rr.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mloftis@wgops.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).