Date: Sat, 31 Jan 1998 18:52:03 -0600 From: Karl Denninger <karl@mcs.net> To: Jonathan Lemon <jlemon@americantv.com> Cc: Brian Tao <taob@nbc.netcom.ca>, current@FreeBSD.ORG Subject: Re: RAID controllers - folks, check this thing out Message-ID: <19980131185203.60841@mcs.net> In-Reply-To: <19980131160423.30536@right.PCS>; from Jonathan Lemon on Sat, Jan 31, 1998 at 04:04:24PM -0600 References: <19980131144604.03410@mcs.net> <Pine.GSO.3.95.980131161942.27817Z-100000@tor-adm1> <19980131155527.19192@mcs.net> <19980131160423.30536@right.PCS>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Jan 31, 1998 at 04:04:24PM -0600, Jonathan Lemon wrote: > On Jan 01, 1998 at 03:55:27PM -0600, Karl Denninger wrote: > > RAID 5, due to the way it stripes parity across the volumes, has a "sweet > > spot" in performance at 5 spindles. > > This is only true if your stripe set is 4 spindles. There's nothing > wrong with using a stripe set of 8 spindles (9 devices), except that > it tends to make small writes slower, since your data is spread out > over more devices. > > 5 devices is not an inherent property of RAID 5, AFAIK. > -- > Jonathan The trade-off, however, between slowing down write performance, parity computation, stripe size, etc seems to be right around 5 spindles. The CMD controller can set up multiple RAID sets however, so if you want to run, say, 10 devices you can set up two RAID 5 sets. You *CAN* use a different number of devices - its just that the way things end up working out, the performance .vs. redundancy tradeoff seems to converge right at 5 devices. -- -- Karl Denninger (karl@MCS.Net)| MCSNet - Serving Chicagoland and Wisconsin http://www.mcs.net/ | T1's from $600 monthly to FULL DS-3 Service | NEW! K56Flex support on ALL modems Voice: [+1 312 803-MCS1 x219]| EXCLUSIVE NEW FEATURE ON ALL PERSONAL ACCOUNTS Fax: [+1 312 803-4929] | *SPAMBLOCK* Technology now included at no cost
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980131185203.60841>