Information

Parallels Automation for the Cloud Infrastructure module automatically calculates Hardware Nodes' capacities upon paci_hn_psbm package installation. However, the default values are set to physical resource capacities and do not allow you to provision more virtual servers onto the Hardware Node if the total amount of allocated resources exceeds these capacities.

Usually, virtual servers do not consume all allocated resources, which is why it is possible to change the pre-calculated capacities.

The following capacities can be edited:

  • CPU Cores
  • CPU Frequency
  • RAM
  • Bandwidth
  • Diskspace

It is possible to modify the automatically defined values to the desired values as per the Provider's Guide. See Configuring Node Resources.

Example:

If the Hardware Node has 32 cores with 3GHz each, then the CPU power of the host is 32x3GHz.

If a customer orders vCPUs (1GHz each), then the overall Node capacity for this specific VM configuration would be 96 vCPUs with 1GHz.

At the same time, no customer can order a VM/CT with more than 32 vCPUs (because that would exceed the real number of cores on the host).

If you want to overcommit (sell more CPU power than is available on the host), you can change the CPU power allocation settings for the Hardware Node:

a. Change the number of CPUs.

b. Change the CPU frequency (MHz). It is recommended that you estimate average resource consumption rate before setting up new capacities, so as to avoid performance issues in the case of massive overselling.

For more information about how CPU power is calculated, refer to these articles:

  • 111354 What does "CPU limit in MHz" mean in Parallels Server Bare Metal 5?

  • 112588 CPU limits design in Parallels Server Bare Metal 5 and Parallels Virtuozzo Containers for Linux

Internal content

Link on internal Article