Tag Archive: qemu


for KVM, using an actual network bridge is desired for performance over NAT. There is a virtual bridge used by XEN (virbr0) that you won’t need.
To disable it:

$ virsh net-destroy default
$ virsh net-undefine default

# If Redhat based distro:
$ service libvirt-bin stop
# if Debian/Ubuntu based distro:
$ service libvirtd restart  

$ ifconfig
 

A quick test (so take it with a grain of salt):

  1. Huge pages is slightly faster than not using huge pages (~10% with 4 winxp virtual machines copying 512MB from one memory location to another).
  2. KSM is slightly slower then not using KSM (~5% with 4 winxp virtual machines copying 512MB from one memory location to another).

So, at first glance it would appear that we can use a loose rule of thumb:

  1. to consolidate the maximum number of machines, use KSM as it will allow you to over commit the amount of memory on your box.
    1. Risk: if the memory pages are significantly different, you may start swapping in a very bad way. This is where monitoring comes in
  2. to give the best performance to a number of machines, use huge memory pages.. does not allow you to over commit
    1. Risk: if you don’t leave enough memory for the host os, you can crash your machine. This is also where monitoring comes in

I installed Ubuntu Eucalyptus cloud (#UEC) on a couple of servers at home. I was naively expecting it to be similar to VMware. Oh I was wrong.

UEC is the Eucalyptus ‘cloud’ software running on Ubuntu servers. The instances would run under the KVM or Xen hypervisors. Ubuntu’s defaulted to running KVM but isn’t restricted by it.

“Eucalyptus is a software available under GPL that helps in creating and managing a private or even a publicly accessible cloud. It provides an EC2 compatible cloud computing platform and S3 compatible cloud storage platform…” Eucalyptus Beginner’s Guide

This is how a vm works in UEC:

You publish an image that you previously installed all the software you need (or you can download such an image). There can be as many instances of this image running as the hardware supports (spread across N number of backend Node Servers).

All instances are transient. As long as an image is running, it will have a physical presence.. but when it is shutdown, the instance is removed including any work performed in the instance.

“static” storage is available via a Storage Server. After you create an image but before you start it, you have the opportunity to attach storage to the instance.

Diagram courtesy of Eucalyptus Beginner’s Guide – UEC edition

While the management tools for UEC, and therefore Amazon EC2/EM, are primitive compared to VMware, it should be able to scale much higher with far less focus on which ‘virtual machine’ goes where.