Cloud Management System(CMS) using the libvirt API

One Paragraph of project description goes here

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

Clone the repository to get a copy for yourself

$ git clone https://github.com/mohapeameya/libvirt-vm-monitor.git

Change current directory to libvirt-vm-monitor

$ cd libvirt-vm-monitor

Install QEMU/libvirt/KVM on your system. Check out the following links for instructions:

1. KVM Installation for Ubuntu
2. How to Install and Configure KVM on Ubuntu 18.04 LTS Server

Install the following dependencies if not already installed:

$ sudo apt install qemu qemu-kvm -y
$ sudo apt install libvirt-daemon-system libvirt-clients libvirt-bin libvirt-dev -y
$ sudo apt install virt-top -y
$ sudo apt install bridge-utils virt-manager -y
$ sudo apt install libosinfo-bin -y
$ sudo apt install libguestfs-tools -y
$ sudo apt install python3 -y
$ sudo apt install python-tk -y
$ pip3 install python-libvirt matplotlib

Installing

Start virt-manager to allow VM installation

$ sudo virt-manager

VM Installation

Install Ubuntu 18.04 LTS Server as Guest VM from ISO image using virt-manager as follows:

1. Select Ubuntu 18.04 LTS Server ISO image during installation
2. Allocate 2 GB RAM, 1 VCPU and 6GB Disk Space
3. Set Name: ubuntu1804-x
   where x is the instance # starting from 0
4. Network: virtual network'default': NAT
5. Network interface
   name: ens3
   type: eth
   IPv4 method: manual
   subnet: 192.168.122.0/24(subnet same as virbr0 interface on host)
   address: 192.168.122.y where y = x + 2 and x is the instance #
   gateway: 192.168.122.1
   name server: 192.168.122.1
6. Use entire disk for installation
7. Set your server's name: ubuntu1804-x
   where x is the instance #
8. Finish installation of the VM and store its xml configuration file in host at ./xmlConfig/ directory with the same naming convention(x is the instance # of the VM)

$ sudo virsh dumpxml ubuntu1804-x > ./xmlConfig/ubuntu1804-x.xml

Server Installation in VM

To test the CMS, run a CPU intensive server in each VM instance as follows:

1. Create /etc/rc.local in the VM if not already present

$ sudo touch /etc/rc.local

2. Edit /etc/rc.local file using vim

$ sudo vim /etc/rc.local

3. If file is newly created, add the line below at start of the file

#!/bin/sh -e 

4. Add the following lines at the eof

IPADDR=`ifconfig ens3 | grep "inet " | awk -F'[: ]+' '{print $3}'`
sudo /root/server $IPADDR 55555 &
exit 0	

5. Give permissions if the file was created

$ sudo chmod +x /etc/rc.local

6. Create server.cpp in /root/

$ sudo touch /root/server.cpp

7. Shutdown VM to enable safe file editing from host

8. Use virt-edit on host to edit server file /root/server.cpp and add content(x is the instance # of the VM):

$ sudo virt-edit -d ubuntu1804-x /root/server.cpp

9. Copy-paste server code into the VM's /root/server.cpp file and save changes

10. Install g++ on the VM

$ sudo apt install g++

11. Compile the server program

$ sudo g++ --std=c++11 -o /root/server /root/server.cpp -lpthread

12. Reboot the VM to enable the server

Repeat VM and Server Installation steps for each instance of VM.

Client/Load-generator application

Compile the client/load-generator application on host

$ g++ --std=c++11 -o loadgen loadgen.cpp -lpthread

Running the test

1. Run the CMS application

$ python3 ./cms.py

2. Run any single VM instance through virt-manager as a default server, say instance 0(static IP 192.168.122.2). If you run multiple instances with no load on the VMs, CMS will shutdown all but 1 instance automatically.
3. Run client/load-generator application on host

$ ./loadgen <default_server_ip> <monitor_ip> <min_threads> <max_threads> <test_time_in_secs|time_per_thread_in_secs>

For example:

$ ./loadgen 192.168.122.2 127.0.0.1 15 50 5

4. As the client/load-generator increases/decreases the load on the VMs, CMS will automatically scale up/down the number of VMs without loss of any requests from the client/load-generator. One can confim auto-scaling behaviour using virt-manager.

5. Graphical visualization of test data

$ python3 ./plot.py

Built With

• libvirt - an open-source API, daemon and management tool for managing platform virtualization

Authors

• Ameya Mohape - Initial work - mohapeameya

License

This project is licensed under the GNU GENERAL PUBLIC LICENSE Version 3 - see the LICENSE file for details