Originally posted by: c_sterling
I'm getting some interesting burn-in numbers from an AMS2100 I've got, and would like to see if anyone as seen anything similar.
Using the same Jetstress config (number and size of databases) I'm seeing overall higher latency when running 'on' a DP pool of 3, 6+1's (striped 'down' the left side of 3 disk shelves) vs a single 7+1 (on one disk shelf).
config is single 4GB qlogic card using targets 0a and 1b (switched, targets at 4Gb) with HDLM (windows 2003, 32bit), disks are 300GB, SAS 15k.
I'm wondering if there's some CPU overhead on the controllers due to DP. Thanks...