Introduction

• This is cluster-programming library that provides the interfaces for creating/using the clusters containing multiple server nodes
• The working is simple and easy to use the abstraction functions
• The cluster configuration is to be specified in a configuration file
• The main function of these library are described below

How to use it?

• Specifying the cluster configuraton with config file:

  • This file contains the details about each server in the cluster

  • The name of this file is as specified by the user

  • This file contains the entries as a pair of server id and IP:port address which are comma separated eg:

      200, tcp://127.0.0.1:1201  
      300, tcp://127.0.0.1:1202  
    

  • The node ids and ip:port addresses for all the servers in a cluster must be unique else there will be a parse error

• This library provides a number of basic functions to use for cluster set-up and managing communications between them. These function roles are explained in the corresponding comments in the code

How to test it?

• Go to the directory where the cluster_test.go file is present

• Set the testing parameters (default values are also set) like when to timeout, how many times to loop, whether to broadcast/unicast messages and display intervals (after how much time to show the current status)

• The functions of these parameters are mentioned in the comments

• To print the details of the tests, (how many nodes in cluster, how many messages, kind of messages broadcast/unicast/both, total sent, total received and total expected messages) set the flag "PRINT_DETAILS" in the cluster_test.go file to true.

• To run, hit the following command:

    $ go test github.com/sagar-sontakke/cluster  
  

References

• Go language tutorial (http://tour.golang.org/#1)
• Go package info (for atomic instructions) (http://golang.org/pkg/sync/atomic/#AddInt64)