I was expecting better performance from my recently deployed ESXi 5 host. I would appreciate any suggestions on performance tweaking and/or an explanation on why a process on any particular VM will be limited (considering the host as a whole) even though the VM configuration does not indicate any resource allocation limit. Right now I am just focusing on CPU performance. I don't have the time or expertise to drill down into the bowels of vSphere; hoping for general guidelines or tweak suggestions from the user community experts as reading the best practices guides does not help much. Or, am I just not understanding the feedback I am seeing in the vSphere client during the test.
Host:
ESXi 5.0 (booting from 2GB SD - redundant)
Dell PowerEdge R520
2 x E5-2430 @ 2.20GHz (12 cores, 24034 MHz total)
48GB RAM
RAID 10 - 6 15K SAS
PassMark CPU benchmark - VM #1:
CPU Mark = 2259, peak 3,500 MHz (for comparison, a single E5-2430 should be closer to 4200)
Windows 2008 R2 x64
2 vCPU, 4GB RAM
Resouces config:
CPU - Shares NORMAL, Reservation 0 MHz, Limit UNLIMITED
Adv CPU - Hyperthreading Core Sharing Mode ANY
PassMark CPU testing - VM #2:
CPU Mark = 4408, peak 6,000 MHz (for comparison, a single E5-2430 is closer to 4200)
Windows 2008 R2 x64
4 vCPU, 8 GB RAM
Resouces config:
CPU - Shares NORMAL, Reservation 0 MHz, Limit UNLIMITED
Adv CPU - Hyperthreading Core Sharing Mode ANY
Note, the CPU benchmark on VM #2 is close to a single CPU PassMark benchmark but during the test it still only consumes approximately 6,000 MHz (peak) of my host's 24,000 MHz when monitored with vSphere client.
Thanks in advance...