Hi,
I need experts recommendation on one of my scenario.
I have two applications (vApp1, vApp2) that needs to be hosted on a virtualized environment. The environment is virtualized using vSphere5 and vCenter. The capacity of the environment is around 100 Logical Cores , 500 GB Memory and 10 TB of Storage. Each vApp has web, app and heavy weight DB servers.
My Scenario is I have to bring up vApp1 scaled up say vApp1 will be occupying 80 Cores, 400 GB while vApp2 will in scale down with 20Cores, 100GB and I will do some executions on vApp1, get the results and then scale down vApp1 to say 20 Cores, 50GB to keep it alive , do minimum stuff while I now scale up vApp2 to 80 Cores, 400GB and do some executions on vApp2 to the results from vApp2.
To achieve that I am using some powercli scripts to modify the resource allocation at vApp and all the VMs levels within the vApp.
My question here is when I adjust the resource allocations for the vApp and VMs say for example DB Server from CPUReservationMhz from 20,000 Mhz to 2000 Mhz , the total number of vCPU count in DB Server of vApp1 remains the same say 10 vCPU. In this case how does the 2000Ghz being served from 10 vCPU , will one logical core will provide the whole 2000 Ghz needed or that is distributed from 10 logical cores?
Also is it recommended to bring down the vCPU count also say to 2 when I scale down the vApp. Similarly I want to know the behavior on the memory side as well.
I need some best practices to be followed on this scenario.
Thanks & Regards
Aravindakumar