vROps "virtual disk" vs. "Disk"

[TheVMinator]

What is the difference between "virtual disk" and "disk" in vROps?  If all the disks on my VM are virtual, and I'm not using any raw device mappings, are these two counters the same?

For example, if I'm using counters like:

disk|totalLatency_average

virtualDisk|totalLatency

why might I choose to use  a "virtual disk" counter instead of a "disk" counter?

Thanks!

[MichaelRyom]

Can give you the answer as I don't know.

But this is the differentiation of disk|totalLatency_average:

The average amount of time taken for a

command from the perspective of a

Guest OS. This is the sum of Kernel

Command Latency and Physical Device

Command Latency.

If you look at the read latency of the virtualdisk one could argue that the differentiation sound like it the same.

virtualDisk|totalReadLatency_average

Read Latency (ms)

Average amount of time for a read

operation from the virtual disk. Total

latency = kernel latency + device latency.

[sxnxr · Accepted solution]

The virtual disk group of performance counters supports real-time performance metrics (such as latency and read- and write-speeds) for I/O operations on virtual disks

The datastore group of performance counters supports real-time performance metrics (such as latency and read- and write-speeds) for I/O operations on datastores.

The disk group of counters support metrics for I/O (input/output) performance (such as latency and read- and write-speeds), and utilization metrics for storage as a finite resource.

Disk-I/O counters support metrics for both physical devices and virtual devices:

• A host reads data from a LUN (logical unit number) associated with the physical storage media.
• A virtual machine reads data from a virtual disk, which is the virtual hardware presented to the Guest OS running on the virtual machine. The virtual disk is a file in VMDK format

I tend to use the top two as they give the most up to date real-time stats (90% of stats are latest) the disks metric are more average, summation