Capacity Planning on

Overview

Mon, 01 Jan 0001 00:00:00 +0000

How Should I Read This Chapter

This Chapter shows the different mechanisms available to administrators and users for controlling the capacity assigned to Virtual Machines:

First, you can control the apparent capacity of Hosts by configuring host overcommitment.

Similarly, you can limit the resources that are made available to users, by using the quota system.

Finally, some workloads may require that you co-allocate or coordinate the capacity assigned to a group of Virtual Machines. Affinity and placement rules can be set for VM groups.

Additionally, you can consult the Scheduler section to fine-tune the scheduling policies that control how resources from Hosts, Datastores, and Virtual Networks are allocated to Virtual Machines.

Host Overcommitment

Mon, 01 Jan 0001 00:00:00 +0000

Before allocating a VM to a Host, the Scheduler checks that the capacity requested by the VM fits in the available capacity of the Host. The overall number of VMs assigned to a Host can be controlled by:

Adjusting the total capacity announced by each Host.

Adjusting the capacity requested by the VM.

Virtual Machine Capacity

The resource allocation of the VM is expressed with the following attributes:

Usage Quotas

Mon, 01 Jan 0001 00:00:00 +0000

The quota system tracks user and group usage of system resources, and allows the system administrator to set limits on the usage of these resources. Quota limits can be set for:

users, to individually limit the usage by a given user.
groups, to limit the overall usage by all the users in a given group. This can be of special interest for the OpenNebula Zones and Virtual Data Center (VDC) components.

Resource Limits

The quota system allows you to track and limit usage on:

VM Groups (Affinity)

Mon, 01 Jan 0001 00:00:00 +0000

In enterprise-grade cloud environments, orchestrating the physical placement of virtualized workloads is paramount to achieving infrastructure resilience, performance optimization, and strict fault isolation. OpenNebula addresses these requirements through its VM Group architectural subsystem (Affinity and Anti-Affinity Groups). This mechanism enables system administrators and cloud architects to define relationship policies among multi-VM application tiers. Operational realities such as urgent infrastructure maintenance, emergency capacity rebalancing, or live patching frequently demand that in certain cases these rules must be temporarily bypassed.