Date: Mon, 6 Nov 2006 15:40:54 -0600 From: "Rick C. Petty" <rick-freebsd@kiwi-computer.com> To: Oles Hnatkevych <don_oles@able.com.ua> Cc: freebsd-perfomance@freebsd.org, freebsd-geom@freebsd.org Subject: Re: geom stripe perfomance question Message-ID: <20061106214054.GB75478@keira.kiwi-computer.com> In-Reply-To: <512766008.20061106211234@able.com.ua> References: <961295086.20061105000919@able.com.ua> <200611061204.kA6C4FXt079703@lurza.secnetix.de> <512766008.20061106211234@able.com.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Nov 06, 2006 at 09:12:34PM +0200, Oles Hnatkevych wrote: > > What really is surprising that disks on the same ATA channel on VIA > controller behave much worse than half of their full speed. Yet on a > different ATA channels they perform flawlessly. Not that surprising at all, given how ATA works. It's even worse on Intel chipsets or particularly on any SATA controllers which put multiple drives on the same channel. Thankfully, most nForce boards have a separate channel for each device. Cheap Intel crap. :-P > Seems like the gmirror(8) man page should contain a separate paragraph > telling about peculiarities using old ATA controllers or using disks > on the same ATA channel. ;-) I'd say this is pretty obvious. When two devices share the same cable, I wouldn't expect the throughput to increase at all, in fact it should decrease, what with protocol overhead and master/slave synchronization. And anyway, I doubt gmirror's the correct place to put this information. Possibly ata(4), since it affects any operations which communicate to two devices on the same channel. -- Rick C. Petty
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061106214054.GB75478>