Talk:RAID, performance tests
From FreeBSDwiki
raw info temporarily moved from main article page here
Equipment:
Athlon X2 5000+ 3x Western Digital 250GB drives (WDC WD2500JS-22NCB1 10.02E02 SATA-300) 2x Western Digital 500GB drives (WDC WD5000AAKS-00YGA0 12.01C02) Nvidia nForce onboard RAID controller, Promise TX2300 RAID controller
Athlon 64 3500+ 5x Seagate 750GB drives (Seagate ST3750640NS 3.AEE SATA-150) Nvidia nForce onboard RAID controller
Procedure:
/usr/bin/time -h measuring simultaneous cp of 3.1GB files to /dev/null files generated with dd if=/dev/random bs=16M count=200 simultaneous cp processes use physically separate files write performance tested with dd if/dev/zero bs=16M count=200 sysctl -w vfs.read_max=128 unless otherwise stated
Notes:
system default of vfs.read_max=8 bonnie++ was flirted with, but couldn't figure out how to make it read big enough chunks of data to ever once hit the disk instead of the cache!
Ancillary data:
write performance (1 process) 5 250GB/500GB, graid3 : 153 MB/s 5 250GB/500GB, graid3 -r : 142 MB/s 1 500GB drive : 72 MB/s 1 WD Raptor 74GB drive (Cygwin) : 60 MB/s 1 250GB drive : 58 MB/s 5 250GB/500GB, gmirror round-robin : 49 MB/s 5 250GB/500GB, gmirror split 128k : 49 MB/s
1 process 5 250GB/500GB, graid3 : 213 MB/s (dips down to 160MB/sec) 5 750GB disks, graid3 : 152 MB/s (wildly fluctuating 120MB/s-200MB/s) 3 250GB disks, graid3 : 114 MB/s (dips down to 90MB/sec) 1 500GB disk : 76 MB/s 1 750GB disk : 65 MB/s (60MB/s-70MB/s) 5 250GB/500GB, gmirror round-robin : 63 MB/s 3 250GB disks, gmirror round-robin : 59 MB/s 1 250GB disk : 56 MB/s (very little variation) 3 250GB disks, gmirror split 128K : 55 MB/s 5 250GB/500GB, gmirror split 128K : 54 MB/s
2 processes 5 250GB/500GB, graid3 : 128 MB/s (peak: 155+ MB/sec) 5 750GB disks, graid3 : 125 MB/s (peak: 140+ MB/sec) 3 250GB disks, graid3 : 98 MB/s (peak: 130+ MB/sec) 3 250GB disks, graid3 -r : 88 MB/s (peak: 120+ MB/sec) 2 250GB disks, nVidia onboard RAID1 : 81 MB/s (peak: 120+ MB/sec) // initial test had flawed data - retested for final article 5 250GB/500GB, gmirror round-robin : 73 MB/s 2 250GB disks, Promise TX2300 RAID1 : 70 MB/s (peak: 100+ MB/sec) // initial test had flawed data - retested for final article 1 500GB disk : 70 MB/s 1 250GB disk : 56 MB/s (peak: 60+ MB/sec) 2 250GB disks, gmirror round-robin : 55 MB/s (peak: 65+ MB/sec) 3 250GB disks, gmirror round-robin : 53 MB/s 5 250GB/500GB, gmirror split 128K : 50 MB/s 3 250GB disks, gmirror split 128K : 46 MB/s
3 processes 5 250GB/500GB, graid3 : 106 MB/s (peak: 130+ MB/sec low: 90+ MB/sec) 5 250GB/500GB, graid3 -r : 103 MB/s (peak: 120+ MB/sec low: 80+ MB/sec) 1 500GB disk : 72 MB/s 5 250GB/500GB, gmirror round-robin : 69 MB/s 1 250GB disk : 55 MB/s 3 250GB disks, gmirror round-robin : 53 MB/s 3 250GB disks, gmirror split 128K : 49 MB/s 5 250GB/500GB, gmirror split 128K : 47 MB/s
4 processes 5 250GB/500GB, graid3 : 105 MB/s (peak: 130+ MB/sec low: 90+ MB/sec) 5 250GB/500GB, graid3 -r : 105 MB/s (peak: 120+ MB/sec low: 80+ MB/sec) 1 500GB disk : 72 MB/s 5 250GB/500GB, gmirror round-robin : 71 MB/s (peak: 75+ MB/sec low: 64+ MB/sec) 3 250GB disks, gmirror round-robin : 65 MB/s 1 250GB disk : 55 MB/s 3 250GB disks, gmirror split 128K : 55 MB/s 5 250GB/500GB, gmirror split 128K : 47 MB/s (peak: 59+ MB/sec low: 31+ MB/sec)
5 processes 5 250GB/500GB, graid3 -r : 107 MB/s (peak: 120+ MB/sec low: 80+ MB/sec) 5 250GB/500GB, graid3 : 105 MB/s (peak: 130+ MB/sec low: 90+ MB/sec) 5 250GB/500GB, gmirror round-robin : 72 MB/s (peak: 80+ MB/sec low: 67+ MB/sec) 1 500GB disk : 72 MB/s 1 250GB disk : 56 MB/s 5 250GB/500GB, gmirror split 128K : 47 MB/s (peak: 60+ MB/sec low: 35+ MB/sec)
vfs.read_max=8, 2 parallel cp processes 3 250GB disks, gmirror round-robin : 31 MB/s 1 250GB disk : 27 MB/s 2 250GB disks, gmirror round-robin : 23 MB/s
Preliminary conclusions:
system default of vfs.read_max=8 is insufficient for ANY configuration, including vanilla single-drive gmirror read performance sucks - Promise read performance sucks - nvidia read performance sucks for single-process graid3 is the clear performance king here and offers very significant write performance increase as well SATA-II seems to offer significant performance increases over SATA-I on large arrays