Hello -
In VCOPS (advanced), under datastore workload there are metrics for Estimated Read Capacity/Write Capacity for IOPs. Does VCOPS take into consideration the underlying type and number of disks within a RAID Group plus the read/write penalty? Just trying to understand how VCOPS does its calculation.
Thank you.
It does IO via observation, as it would not have the ability via ESXi to see the underly disk architecture (RAID 5/6/10/etc or SATA vs SAS) It just reports the data seein in vCenter.
However, once a metci is normalized over time, we would track if it goes out of the expected area (Dynamic Threshold) or other values that matter for a Datastore.
It does IO via observation, as it would not have the ability via ESXi to see the underly disk architecture (RAID 5/6/10/etc or SATA vs SAS) It just reports the data seein in vCenter.
However, once a metci is normalized over time, we would track if it goes out of the expected area (Dynamic Threshold) or other values that matter for a Datastore.
So it just looks at the IO max at any given time and then uses that number as the max capacity?
For disk at the host level it is looking for the max from the host? And for cluster it is adding up all of the host maxs? From the datastore view would it traverse all clusters that are accessing the datastore?
Thanks,
-Matt
Yes ... per host, it does a high water mark.
As for the Datastore level, I'm not familiar with how the data is aggregated laterally or vertically.