RHEV: best practices for reducing the load on the SPM

Solution Verified - Updated -

Environment

Red Hat Enterprise Virtualization (RHEV) 3

Issue

Looking for best practices for reducing the load on the SPM node.

Resolution

1. Set up the right Load Balance policy for your clusters. Evenly Distributed policy is recommended.

2. Increase the value of SpmVCpuConsumption config option. Default is 1.

The way that RHEVM chooses the host to start a VM on is by first defining all candidates according to the following: number of hosts in cluster, host has a status of up, have enough memory to run the VM, have at least number of core as the VM, and swap usage is according to guest config. By adding dummy vcpu consumption tor the SPM host, we will limit the number of VMs running on it.

What is the recommended value for this? Increase the SpmVCpuConsumption by 50% of the number of VM's. If the environment has 30 VM's, increase it by 15.

# rhevm-config -s SpmVCpuConsumption=15
# service ovirt-engine restart

3. Decrease the value of OvfItemsCountPerUpdate
This will reduce the amount of OVF files sent by an update, which is an IO operation writing to the storage, and can cause to SPM overload as well.
NOTE: this value is exposed to config starting RHEV 3.5

# rhevm-config -s OvfItemsCountPerUpdate=10
# service ovirt-engine restart

This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.

Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.