Skip to content

codership/benchmarks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

d1ae0ab · May 3, 2014

History

31 Commits
Dec 28, 2013
Feb 17, 2014
Apr 23, 2014
Jan 24, 2014
Dec 27, 2013
Dec 27, 2013
Apr 23, 2014
Nov 24, 2013
Dec 26, 2013
Jan 3, 2014

Repository files navigation

benchmarks

repository for benchmark scripts and such

Idea

This is (going) to be a set of scripts helping to build a test environment and test your Galera setups. We are using ansible to automate and distribute tasks. You don't need to log on any of the nodes as you are going to manage your cluster/testing from you local machine. So this repo is going to be your working directory for your ongoing tasks.

There are 4/5 modules planed

  • Provisioning (not implemented)
  • Installation (ongoing)
  • Testing (ongoing)
  • Visualising (not implemented)

Provisioning

This module is going to help you to launch instances in AWS or an OpenStack environment.

Installation

This module/scripts are going to install and configure you a

  • Galera cluster
  • Deploys the testing script on a separate machine

Now (20131226) you are going to have a cluster of $yournumber of galera nodes with $yournumber of testing/sysbench nodes pointing to one of the galera nodes.

We are going to support the Debian(Ubuntu) and RedHat OS family. Regarding Galera you can choose between Codership, MariaDB, Percona. ( Right now only Codership is supported on RehHat)

Testing

While you can install your testing hosts. We use sysbench 0.5 for the testing. You can invoke the script standalone or just run a (so called) play. Starting a play/test run you can configure your galera settings. A rolling restart is done for you automatically and result is stored in your /tmp directory

Format

The result is written in yaml and include

  • SHOW GLOBAL VARIABLES
  • SHOW GLOBAL STATUS (before and after each sysbench test)
  • sysbench results It is planned to put to put the environment (virtualisation and (ansible -m setup)) informations into that yaml to. (Maybe we switch to json also) The idea is to have as much informations about a run as we need even for later investigation.

Visualising

This is about extracting and visualise the informations. BIG TODO

WARNING

This is work in progress. It is going to be changed on a daily basis :)

About

repository for benchmark scripts and such

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages