Hi community
Thanks GS gs_khalsa for your help
I am still did not understand how can I put a limit for virtual compute not to exceed the physical compute resources for exempel I do not want to vCPU more than pCPU in the event of host failure.
Lets assume I have 3 hosts with 8 pCPU per host. each host allocated 3 VMs with 2VPCU so there is no over subscription per host which is that what I wish to have
I case of host failed one host of the two will be over provisioned with vCPU more than pCPU. How can I avoid this I do not mind to the last VM which will cause the over superscription to be off and not to start in the new host.
Thanks
Reservations will ensure that VMs get the resources assigned to them. VMs without reservations will start and get whatever is left. If they have HA admissions control enabled, HA will not allow VMs to start if the host doesn't have resources to meet the reservations.
If you have more questions I'd highly recommend checking this: Introduction | vSphere 6.x HA Deepdive out first.
Set CPU reservations on the VMs that you want to run and not on the ones you don't.
Thanks alot GS ,
Does CPU reservation prevent any CPU overcommitment ?
Those VMs without any CPU reservation in host failure will not restart in case of CPU overcommitment ?
Reservations will ensure that VMs get the resources assigned to them. VMs without reservations will start and get whatever is left. If they have HA admissions control enabled, HA will not allow VMs to start if the host doesn't have resources to meet the reservations.
If you have more questions I'd highly recommend checking this: Introduction | vSphere 6.x HA Deepdive out first.