ESX 3.0.1 running in an 8-way (4x8218) DL585G2 with 32G with a QLA200 HBA. Guest machine is
running Novell's SLES 10 with 1G memory. Disk is coming from a Nexsan SATABeast configured
for RAID6 and using a 2Gbps Sanbox 5202 switch.
It's not awesome, a gazillion times better than using the local SCSI RAID on the DL585.
SLES10 users, I used bonnie++ 1.01d rpm's from the SUSE 10.1 Professional.
ESX server was lightly loaded (about 10 guests)... might have thrown off the numbers smidge.
Version 1.01d ------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 eightcore-sles10 2G 76755 95 149797 53 61544 29 96998 83 167702 23 302.6 0 ------Sequential Create------ --------Random Create-------- -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete-- files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP 16 +++++ + +++++ + 29948 99 +++++ + +++++ + 27773 99 eightcore-sles10,2G,76755,95,149797,53,61544,29,96998,83,167702,23,302.6,0,16,+++++,+,+++++,+,29948,99,+++++,+,+++++,+,27773,99
Oh.. the virtual disks for the guest are off of vmfs which in on the SAN storage. In case some were wondering. We placed a LUN from the SATABeast under vmfs control.
Did you saw this thread ?
Thanks... I may run iometer and post to that thread. Let's see if some other can post some bonnie++ benchies here.
Ick. iometer may be open source... but it's really, really, really bad code. I don't think I want to try it on Linux.