What is the maximun number of Vms deployed per host?

Community Support
1

Hello.

I want to create a private cloud using sunstone consisted by: a opennebula server and a working node (host), on which several virtual machines will be instantiated.Currently I am using the following releases:

Opennebula: version 4.12.1
Hypervisor: kvm
Host:
Memory: 7.7GiB
OS: Ubuntu 14.04 64-bit
Disk: 199.4 GB
Images: ttylinux-kvm from opennebula marketplace at http://marketplace.c12g.com/appliance

The images are stored inside a shared filesystem datastore on the host machine.

Currently, when using either persistent or non-persistent images, everything works smoothly when I
deploy up to 3 VMs. However, when I try to deploy a fourth virtual machine it cannot be deployed (remains in pending state), until I either delete it or undeploy one of the active ones.

At the dashboard of the VM which is in pending state I get the following message at the placement section:

No host meets SCHED_REQUIREMENTS: !(PUBLIC_CLOUD = YES)

Has anybody had the same problem before? Is there a configuration parameter on the host (either via sunstone or via the terminal) that limits the number of VMs to 3? Could anyone explain why this is happening?

Any help would be highly appreciated.

Thank you very much for your time.

2

Hi,

You may want to take a look to the scheduler documentation:
http://docs.opennebula.org/4.14/administration/references/schg.html

The scheduler will try to deploy VMs in pending state in a host with enough capacity and that meets the requirements defined in the VM template.

You should check the allocated values of the host (CPU & MEMORY). Probably, you are trying to instantiate the ttylinux template using CPU=1 therefore you are limiting the number of VMs to the number of cores of the host. You can define a number lower than 1 to avoid that, for example CPU=0.1 in the VM template

3

Hello.

It was indeed the “CPU=1” at the template that was causing the problem. After following your suggestion, I can now deploy more than three VMs with no problems. Therefore the issue has been resolved.

Thank you.