Date: Sat, 24 Jan 2004 19:48:15 -0800 From: Rishi Chopra <rchopra@cal.berkeley.edu> To: Charles Swiger <cswiger@mac.com> Cc: questions@freebsd.org Subject: Re: Adaptec 2400A Performance Message-ID: <40133C7F.2090803@cal.berkeley.edu> In-Reply-To: <A35EEBF2-4E97-11D8-9B19-003065A20588@mac.com> References: <401207EF.7030005@cal.berkeley.edu> <A35EEBF2-4E97-11D8-9B19-003065A20588@mac.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Just wanted to make sure I wasn't skipping any obvious steps. One other (slightly lamer) question: if my device configures as da0, is that scsi or ide? The reason I ask is I wish to write a custom kernel, and would like to eliminate all unnecessary configurations/devices. Charles Swiger wrote: > On Jan 24, 2004, at 12:51 AM, Rishi Chopra wrote: > >> I was rather disappointed with the results. Can anyone suggest what >> might be causing such slow disk speeds, or whether these speeds are >> out of the ordinary for a 4-disk FreeBSD RAID5 installation? I have >> done nothing to configure the card aside from striping the array in >> BIOS; FreeBSD seems to automatically detect the disks. > > > For us to be able to comment beyond generalizations, it's necessary to > also benchmark how a single disk performs. I can still answer your > question, though: > > RAID-5 is slow. RAID-5 trades availability against performance and > hardware costs. With RAID-0, n drives gives n drives' worth of usable > space. With RAID-5, n drives gives n-1 drives' worth of usable space. > The performance is between RAID-0 and RAID-1 is comparible for large > accesses. For small accesses, particularly small writes, RAID-5 > performance is much worse than plain RAID-0 or a plain disk. > -- Rishi Chopra http://www.ocf.berkeley.edu/~rchopra
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?40133C7F.2090803>