Date: Fri, 04 May 2007 03:42:52 +0200 From: Fluffles <etc@fluffles.net> To: "Wilkinson, Alex" <alex.wilkinson@dsto.defence.gov.au> Cc: freebsd-geom@freebsd.org Subject: Re: graid5 compilation problem Message-ID: <463A8F9C.8020205@fluffles.net> In-Reply-To: <20070504011646.GD15396@obelix.dsto.defence.gov.au> References: <1847971490.20070503223307@uzvik.kiev.ua> <971656.83802.qm@web30315.mail.mud.yahoo.com> <593515236.20070504022900@uzvik.kiev.ua> <463A7558.9090600@fluffles.net> <20070504011646.GD15396@obelix.dsto.defence.gov.au>
next in thread | previous in thread | raw e-mail | index | archive | help
Wilkinson, Alex wrote: > 0n Fri, May 04, 2007 at 01:50:48AM +0200, Fluffles wrote: > > >deniswork@uzvik.kiev.ua wrote: > >> Thanks, this help! > > Question: Is there a graid5 guide anywhere ? > I would like to test this code out. > I'm working on one; but basically it's compile (works on both 6 and 7-CURR) and go using "graid5 label <volumename> <disks>". and you can use it straight away. I've done extensive benchmarking but results depend on: - hardware, especially CPU/MEM/BUS technology - disk bus bandwidth; avoid PCI whenever possible, PCI is depricated. Use onboard connectors directly tied to chipset if possible - the version of graid5, stable or TNG (take the graid5-eff.tbz package instead of graid5.tbz); TNG has nice speed improvements - disk speeds themselves Under the right conditions (dualcore with 8 SATA disks on embedded bus) it is possible to achieve 400MB/s *write* throughput, which is short of amazing considering this is 100% software RAID, and leaves many XOR-hardware RAID solutions behind. ;-) - Veronica
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?463A8F9C.8020205>