Hi, What I have to say probably will be of very little help, but anyway: We have here an IBM NAS 200, which has a ServeRaid 4H + 6 7200 rpm 72GB disks. It came as a raid5 (+hotswap) and I still did not try other confs. It came with 3 logical disks, of sizes 6GB, 12GB and 260GB. I ran zcav on the third one several times. On the first times it had many points in which it went down to between 10MB/s and 45MB/s (which was the maximum, and it was close to it most of the time). On the last runs it never went below 42MB/s.
I can try to make a stripe/mirror/other stuff on it - it's not yet production. The short answer: I guess 45MB/s is the maximum of each single disk, and raid5 did not give better results on the maximum. Are you sure a stripe of two should give twice the speed of one? Do the specs say so? Did/Can you try Linux's software raid? If you do, you might get better results by connecting one drive to each card (happened to me with IDE - yes, I know SCSI is way faster/more efficient/more expensive). Did your disks gave (individually) constant throuput, or (as is common) faster in the beginning and slower in the end? Did the raid at least made that better (as I think it did for me)? I don't mind giving the graphs, but I do not think it's allowed to publish them as a formal benchmark (even more so considering it was a donation of IBM). About the 32/33/64/66: I am not sure the card works the same under all conditions (except the PCI bottleneck). Maybe it also works slower in the communication to the disks, when running under 32/33? Does the spec say anything about that? BTW, I don't know HD-Tach and Sandra, but zcav+gnuplot give me very nice graphs (at least for my taste). Try: $ zcav /dev/xxx > somefile $ gnuplot gnuplot> plot "somefile" You can also try to run two zcavs on them (without raid). Both on the same card and on different ones. I would love to see the results (and don't mind giving mine, with those IDE disks). -- Didi ================================================================= To unsubscribe, send mail to [EMAIL PROTECTED] with the word "unsubscribe" in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
