On Thursday 20 December 2007 20:10:12 Tracy R Reed wrote: > Dexter Filmore wrote: > > ZFS is software. Don't know how ZFS performs, but being software I *bet* > > it won't write 400MB/s like a hardware controller does. (Ok, 400 at a > > sufficient high disk number, 6 or 8). I got soft raid 5 on my server and > > besides that it is troublesome in terms of stability (had at least 6 > > occurances where a disk flew from the array for no apparent reason, disk > > checks turned out disk was fine so I re-added it each time) it does > > little more then 10MB/s writes at cost of many cpu cycles. > > Your first statement makes no sense. That hardware controller runs > software too. There is no such thing as "hardware is better than > software" or vice versa. They are apples and oranges. You cannot have > one without the other. And RAID 5 sucks in general for performance. On a > P4 system, XOR computation can be performed by the MMX unit > (independently of the main cpu) at 2Gbytes/sec, well above the needs of > I/O systems. So the RAID calculations themselves aren't likely to be > causing noticeable CPU performance differences. In every case that I > have had a machine kicking out disks from a Linux software RAID it was > because there was some issue with either the hardware or the hardware > driver. Some of them have even been intermittent issues where I could > put the disk right back in and it would work fine.
Worked here fine, too, but it puzzled me every time why the disk flew out anyway. dmesg didn't show anything, drives appeared fine. About "there is no hardware raid": even if a P4 can do 10GB/s XOR ops, the system has to talk to the drives via the bus system, feth the data, put it back on the bus and have the controller tell where to write it. Hardware controller does it pretty much without talking to the CPU since it got is own. All the way from retrieving the data from the disks, shifting them to the CPU and then return them is a software thing, no matter how much MMX units there are. Plus, there's a bottleneck unless your controller sits in an PCIe 4x slot. Even then the above stated takes effect. -- -----BEGIN GEEK CODE BLOCK----- Version: 3.12 GCS d--(+)@ s-:+ a- C++++ UL++ P+>++ L+++>++++ E-- W++ N o? K- w--(---) !O M+ V- PS+ PE Y++ PGP t++(---)@ 5 X+(++) R+(++) tv--(+)@ b++(+++) DI+++ D- G++ e* h>++ r* y? ------END GEEK CODE BLOCK------ http://www.stop1984.com http://www.againsttcpa.com -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
