Installing CoreCluster

 

This document will guide you through whole installation process of CoreCluster cloud. You should know basics of the Linux administration and package installation to successfully go through this guide.

System requirements

For minimal installation in non-production environments you will need computer with at least 1GB of RAM and dual core processor. However best experience you will get with at least 4GB of RAM.

Supported operating systems are:

The same requirements are for Computing Nodes, however in this case you should keep in mind to leave at least 1GB of RAM and 1CPU for Node's operating system. If node is shared as workstation (with desktop), reserve more CPU and RAM for the user's applications.

Important! Don't create user with cloudover login! This name is reserved for the system account!

Prepare the system - Management Node and Computing Nodes

Whole installation is done through our public .deb package repository. At this time we don't support RPM-based distributions. Open in your favorite text editor file /etc/apt/sources.list.d/cloudover.list and add one of following lines, matching your distribution:

Debian:

deb http://packages.cloudover.org/debian jessie main

Ubuntu and Linux Mint:

deb http://packages.cloudover.org/ubuntu xenial main
deb http://packages.cloudover.org/nightly nightly main

Next, add CloudOver's public key to apt:

wget http://packages.cloudover.org/cloudover.key -O - | sudo apt-key add -

Package installation

Management Node

Update repository with following command:

sudo apt-get update

When package lists are up to date you can install main CoreCluster package, together with Mysql server and other useful packages:

sudo apt-get install corecluster htop git tcpdump mysql-server

During installation you will be asked for Mysql administrator's password

Computing nodes

Installing computing node is quite similar. Update package lists with:

sudo apt-get update

When package lists are up to date you can install main CoreNode package, together with other useful packages:

sudo apt-get install corenode htop tcpdump

Post installation steps

Networking setup

Corenetwork package is responsible for configurations of network part of whole cluster and the some configurations of operating system. This package should be installed automatically, as dependency on each compute node and management node. Especially you should take care about following settings:

Edit the /etc/corenetwork/config.py file and update following variables on each computing node and management node:

Without updates in network configs, the networking might fail. It causes problems with connectivity to the Internet, inter VM communications and problems with CloudInit

Database connection

By default CoreCluster uses sqlite3 file as the database. You should not use it even for tests. Edit with your favorite editor /etc/corecluster/config.py file and look for the DATABASE variable. This file has Python-like syntax. You should find two databases:

Enter new credentials to database into this file. Don't use root's account! Next, connect with Mysql server by:

mysql -u root -p

You will be asked for root's password chosen during installation. In Mysql shell enter following to create new database (as you entered into config file) and user account:

CREATE DATABASE cloudover;
GRANT ALL PRIVILEGES ON cloudover.* to cloudover@'localhost' IDENTIFIED BY 'your_cloudover_db_password';

If you use Mysql as the logs database, create second database in the same way. It is recommended to use different database with different user. Migrate database with commands:

sudo cc-admin migrate
sudo cc-admin migrate --database=logs
Logs database for nodes

If you want to store logs cluster-wide, in one common database, from all Computing Nodes, then edit at each node following file: /etc/corenode/config.py. Find DATABASE connection details and update logs database. It will be necessary to grant privileges to this database to all hosts ('%' instead of 'localhost'). for example:

CREATE DATABASE cloudover_logs;
GRANT ALL PRIVILEGES ON cloudover_logs.* to cloudover_logs@'%' IDENTIFIED BY 'your_cloudover_LOG_password';
and finally, flush privileges:
FLUSH PRIVILEGES
Create system administrator

Create CoreCluster system administrator with command:

sudo cc-admin createsuperuser

You will be asked for account name and password.

Setting default Computing Node capabilities

By default all computing nodes will find your CoreCluster Management with Autodiscover tools. This makes installation and configuration much faster. To change default values of Computing Node capabilities, edit file /etc/corenode/config.py and update proper values related to cpu, memory and disk capabilities. This values describe maximum amount of resources shared with Cloud. However cloud administrator will be able later to manually override this values in CoreCluster Management database.

Restart your systems

To provide full functionality, restart your CoreCluster Management machine first. After reboot restart all Computing Nodes. This will reload configurations for UWSGI server and all agents across the cloud. It will also make possible to register Computing Nodes in CoreCluster Management database

What next?

Now, you have at least one node and management machine. Go to Core UI installation guide to install graphical web interface, and then to the Administrators guide to add basic resources to your cloud. Also you can add some extensions before going ahead.
< Go back     Author: Maciej Nabozny Published: Sept. 23, 2016, 5:46 p.m.