On Fri, Aug 11, 2000 at 04:34:48PM -0500, Jim Moechnig wrote:
Depends on the application, of course, but I've been surprisedmany times in the past when I thought for sure the Netapp would not be able to keep up. I have a 4x450-MHz E420R with a VxVM RAID-0 device, spread over 16 50GB 7200 rpm drives on two U2SCSI buses. The server also has a Gigabit Ethernet connection to an F740 with one shelf of 36GB 10000 rpm drives (5 data, 1 parity, 1 spare). The local filesystem is vxfs, mounted with delaylog and the largest allowable log area.
If I'm reading this correctly, you're using all software raid for the DAS side, correct? I would be very interested in seeing a similar comparison with a hardware raid controller, as this would offload all the raid overhead from the host cpus to a subsystem better optimized to handle it. This would also unfortunately introduce a lot more variables of the sort Alan Yoder pointed out, as I'm sure the performance under load of the various available raid controllers is far from uniform.
I don't have hard numbers, but when we moved our Oracle database from direct-attached A1000 (dedicated diff. SCSI controler) to a filer we saw a large join that used to take 3 hours drop to 20 minutes, and overall database performance increase perceptably. We're very happy with the Filer's ability to handle large numbers of simultaneous reads/writes. Your milage may vary if you're doing single-threaded reading or writing.
Certainly key is using GigE to connect the Filer to the server. Also, keep in mind that the more disks you put in the Oracle volume, the better your performance. We almost made the mistake of breaking our volumes up into 3-disk sets until our NetApp OEM told us what the performance impact would be.