Hello list,
I am experimenting with an Adaptec raid controler 3410s
which uses the i2o drivers (dpt_i2o or i2o_block). The
machine has both an onboard aic7xxx scsi and the 3410s so I
benchmarked the array using both. However, it seems
softraid is performing much better than the hardware
solution. Before I toss the hardware, I'm wondering if
there is anything I can do to bring the performance up to
that of software raid. The numbers seem a bit low for what
the 3410s is supposed to be capable of.
Details:
Running 4 Seagate 15k rpm 36gb drives in a raid5
configuration. Using hardware raid, these are split 2,2
over the two internal channels of the i2o card. Same with
softraid, the onboard scsi has 2 channels and the drives
are split 2,2. The raid5 set is created with no spares and
a 64kb chunk size in both configurations. Raid set is
optimal before the test is run.
Bonnie results:
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
i2o-raid5-dfl 4G 26229 86 36087 21 19096 8 24737 71 73774 12 573.7 1
soft-raid5 4G 29873 98 75664 45 32513 17 33146 97 135790 40 617.5 2
I am using kernel 2.6.10 with i2o_core and i2o_block for
the hardware raid. I also tried dpt_i2o with no
significant difference.
Best regards,
Shane
--
Shane Wegner
http://www.cm.nu/~shane/
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
