[Pc_Support] 3Ware 7410 v 7500-4 RAID 5
Jason Boxman
jasonb at edseek.com
Fri Mar 18 00:15:43 EST 2005
I just upgraded to a 7500-4, running on a slot by itself on a 64/33 bus. I'm
getting some strange numbers from it using bonnie++. The configuration is
RAID 5 using 64K stripe with 4 x 120GB WD 8MB/cache disks. The io scheduler
used was cfq. The box was idle in single user mode for all tests. The disk
array is presently 60% full and running SGI's XFS filesystem. (The OS is
Linux.) The readahead has been tuned to 1536 * 512 byte sectors, as this
acheives the best `hdparm -Tt` performance over the default of 256.
I'm curious if these results are within reasonable expectations. The initial
55MB/s RAID 5 write was nice, but that isn't consistent. The reads seem a
bit slow, too. Perhaps not near the end of the disk platter, though.
The first test, using a 3GB file size and a 64K chunk size. The kernel was
patched with -ck2 which I forgot to change to sys.vm.mapped=33 from the
default of 66. The former is supposedly more ideal for servers. You'll note
the somewhat better writes over the 7410 benchmarks that follow below. The
reads are essentially unchanged.
nebula:~# cat 2.6.11-ck2-nfs4-7500.out | bon_csv2txt
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
nebula 3G:64k 55041 38 17120 15 49181 19 108.6 3
nebula 3G:64k 38089 26 16991 14 49805 19 108.1 3
nebula 3G:64k 37913 26 16511 14 51549 19 87.9 3
nebula 3G:64k 31405 22 16285 14 49670 19 108.3 3
nebula 3G:64k 34692 24 16726 14 53287 20 84.5 2
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
nebula 16 762 16 +++++ +++ 1520 22 655 15 +++++ +++ 114 1
nebula 16 814 17 +++++ +++ 1704 26 693 15 +++++ +++ 1177 22
nebula 16 808 17 +++++ +++ 1670 28 714 15 +++++ +++ 1108 20
nebula 16 1292 25 +++++ +++ 1718 25 867 19 +++++ +++ 1359 25
nebula 16 1237 27 +++++ +++ 1498 23 757 16 +++++ +++ 1033 19
Now, here's the 7410 using the same kernel and array configuration.
nebula:~# cat 2.6.11-ck2-nfs4.out | bon_csv2txt
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
nebula 3G:64k 44943 31 18851 16 47983 19 99.1 3
nebula 3G:64k 30633 20 19846 17 48842 18 97.6 3
nebula 3G:64k 34129 24 20130 17 51255 19 90.2 3
nebula 3G:64k 27469 19 19205 17 48829 19 98.1 3
nebula 3G:64k 33236 22 17465 15 50421 19 82.3 2
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
nebula 16 706 16 +++++ +++ 892 14 622 14 +++++ +++ 530 9
nebula 16 845 17 +++++ +++ 1035 16 656 15 +++++ +++ 607 10
nebula 16 783 17 +++++ +++ 816 12 650 14 +++++ +++ 585 10
nebula 16 781 17 +++++ +++ 674 9 717 16 +++++ +++ 537 9
nebula 16 679 15 +++++ +++ 826 13 608 13 +++++ +++ 445 7
Finally, the 7410 under the 2.6.10 kernel with -ac10 patchset. The
performance is far more consistent from pass to pass. I'm going to retest
the 7500-4 with 2.6.10-ac10 later.
nebula:~# cat 2.6.10-ac10.out | bon_csv2txt
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
nebula 3G:64k 28435 21 18514 17 60699 25 136.2 4
nebula 3G:64k 25227 21 18478 16 43567 18 151.0 5
nebula 3G:64k 26212 20 18172 16 49778 20 131.5 4
nebula 3G:64k 24929 20 18732 17 53058 22 151.6 5
nebula 3G:64k 24779 20 18752 17 49625 21 170.3 6
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
nebula 16 720 16 +++++ +++ 915 14 618 14 +++++ +++ 461 7
nebula 16 707 16 +++++ +++ 673 10 714 16 +++++ +++ 409 7
nebula 16 707 16 +++++ +++ 783 11 670 15 +++++ +++ 635 11
nebula 16 831 17 +++++ +++ 682 10 763 17 +++++ +++ 448 7
nebula 16 759 17 +++++ +++ 918 14 657 15 +++++ +++ 448 8
--
Jason Boxman
Perl Programmer / *NIX Systems Administrator
Shimberg Center for Affordable Housing | University of Florida
http://edseek.com/ - Linux and FOSS stuff
More information about the Pc_support
mailing list