Date: Mon, 23 Aug 1999 10:12:37 +0930 From: Greg Lehey <grog@lemis.com> To: James Brown <jb@syndicate.net> Cc: freebsd-fs@FreeBSD.ORG Subject: Re: ccd striped array slower than single disk! Message-ID: <19990823101237.C83273@freebie.lemis.com> In-Reply-To: <37C0951B.837EA807@syndicate.net>; from James Brown on Sun, Aug 22, 1999 at 05:26:03PM -0700 References: <37C0951B.837EA807@syndicate.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sunday, 22 August 1999 at 17:26:03 -0700, James Brown wrote: > hello, > > bonnie reports 5636 K/s when using a single disk without ccd. Bonnie lies. > i've been testing ccd performance, using various interleaves. the > best performance i've seen striping with ccd has been 5571 K/s using > an interleave of 32768. Do you mean 32768 blocks or bytes? One's rather small, the other is unnecessarily large. I've found stripe sizes in the order of 256 kB to be optimum. > these are identical disks (maxtor 27.2 gb) and are on two > separate ultra dma/33 busses (one is bus 1 slave, other is bus 2 > master but that shouldn't matter, right?). shouldn't i be seeing > something closer to 9000 or 10000 K/s? That depends on what you're doing. Remember that bonnie uses buffer cache, so the results are usually inaccurate and difficult to repeat. You also don't say which of the bonnie tests you're quoting. Try rawio (in the Ports Collection) instead, and remember that it's usually the random access tests which count. If you're still unhappy, I'd be interested to see how you fare with Vinum instead of ccd. Greg -- See complete headers for address, home page and phone numbers finger grog@lemis.com for PGP public key To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990823101237.C83273>