Date: Mon, 25 Oct 2004 15:25:32 -0600 From: Scott Long <scottl@freebsd.org> To: Brad Knowles <brad@stop.mail-abuse.org> Cc: freebsd-current@freebsd.org Subject: Re: FreeBSD 5.3b7and poor ata performance Message-ID: <417D6F4C.9000404@freebsd.org> In-Reply-To: <p0600205fbda318006656@[10.0.1.3]> References: <14479.1098695558@critter.freebsd.dk><417D25E8.6080804@ng.fadesa .es> <200410251928.01536.victor@alf.dyndns.ws><200410251837.58257.Thoma s.Sparrev ohn@btinternet.com><417D3F12.20302@DeepCore.dk> <417D40A1.9030802@ng.fadesa.es><417D45F1.9090504@freebsd.org> <77F3FD4D-26BE-11D9-9A2F-003065ABFD92@mac.com><417D58B6.5030509@fr eebsd.org > <F5F15CA0-26C5-11D9-9A2F-003065ABFD92@mac.com> <417D65F1.2040809@freebsd.org> <p0600205fbda318006656@[10.0.1.3]>
next in thread | previous in thread | raw e-mail | index | archive | help
Brad Knowles wrote: > At 2:45 PM -0600 2004-10-25, Scott Long wrote: > >> RAID-0 yes, RAID-10 no, at least not for software RAID. The machine >> winds up having to transfer the same data twice across the PCI bus, >> twice through the controller, etc. If the controller is on a simple >> PCI-32/33 bus then it will quickly become saturated. > > > I can't speak for GEOM, but my experience with vinum is that > software RAID, especially RAID-1+0 will totally smoke the bare hardware, > and will smoke most any PCI hardware RAID controller you put it up against. > > Granted, my experience is a few years old, so if you've got any more > recent data that shows in detail what the current situation is and > explains why, I'd love to see it. Until then, I stand by my results at > <http://www.shub-internet.org/brad/FreeBSD/vinum.html>. > Software RAID does indeed have less latency than PCI RAID, and will often result in higher performance. But as was said, there is always a performance vs. reliability tradeoff. And when you are talking about RAID-10 with a bunch of disks, you will indeed start seeing bottlenecks in the bus. Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?417D6F4C.9000404>