Jump to: navigation, search

Difference between revisions of "Rally"

(Why Rally?)
(Replaced content with " The page has been moved to https://rally.readthedocs.io")
 
(120 intermediate revisions by 10 users not shown)
Line 1: Line 1:
== Why Rally? ==
 
  
If you are here, you are probably familiar with OpenStack and know that it's really huge, very huge  ecosystem of co-working services and when something went wrong, or work slow or doesn't scale it's really hard to answer on questions "why", "what" and "where"? Another reason why you could be here is that you would like to build OpenStack CI/CD system that will allow you to improve SLA, performance and stability of OpenStack continuously.
+
The page has been moved to https://rally.readthedocs.io
 
 
Currently OpenStack community is to busy around adding new features and working on CI/CD that ensures that  new patches don't brake specific single node installation of OpenStack. So we would like to help them and add tool that simplify all steps of benchmarking OpenStack at Scale (even in case when you don't have thousands servers). Idea is simple, combine all existing tools to: automate & simplify all steps: 1) Deploying OpenStack somewhere 2) Verify OpenStack installation 3) Benchmark & Profiling it 4) process results & generation reports .
 
 
Before diving deep in Rally arch let's take a look at 3 major high level Rally Use Cases:
 
[[File:Rally-UseCases.png|850px]]
 
 
 
== Use Cases ==
 
 
 
# Investigate how different deployments affect OS performance:
 
#* Find the set of good OpenStack deployment architectures,
 
#* Create deployment specifications for different loads (amount of controllers, swift nodes, etc.).
 
# Automate search for hardware best suited for particular OpenStack cloud.
 
# Automate production cloud specification generation:
 
#* Determine terminal loads for basic cloud operations: VM start & stop, Block Device create/destroy & various OpenStack API methods.
 
#* Check performance of basic cloud operations in case of different loads.
 
# Automate measuring & profiling focused on how new code changes affect OS performance.
 
# Using Rally profiler to detect scaling & performance issues.
 
 
 
 
 
== Architecture ==
 
 
 
Rally consists of 4 main components:
 
 
 
# Deploy Engine, which is responsible for processing and deploying VM images (using DevStack or FUEL according to user’s preferences). The engine is capable of:
 
#* deploying OS on already existing VMs,
 
#* starting VMs from a VM image with pre-installed OS and OpenStack,
 
#* deploying multiple VMs each of which has running OpenStack compute node based on a VM image.
 
# Server Provider, which provides servers (virtual servers) to deploy OpenStack.
 
# Benchmarking Tool, which does benchmarking in several stages:
 
#* runs Tempest tests, reduced to 5-minute length (to save the usually expensive computing time),
 
#* runs a set of benchmark scenarios (using the Rally testing framework),
 
#* collects all test' results and processes them with Zipkin tracer,
 
#* puts together a benchmarking report and stores it on the machine Rally was lauched on.
 
# Orchestrator, which is the central component of the system. It uses the Deploy Engine to run control and compute nodes and to launch an OpenStack distribution and, after that, calls the Benchmarking Tool to start the benchmarking process.
 
 
 
 
 
To dive deeper into Rally architecture, see [[Rally/ArchitectureForDevelopers|Rally architecture for developers]].
 
 
 
 
 
 
 
== Rally in action ==
 
 
 
=== How amqp_rpc_single_reply_queue affects performance ===
 
 
 
To show Rally's capabilities and potential we used ''NovaServers.boot_and_destroy'' scenario to see how ''amqp_rpc_single_reply_queue'' option affects VM bootup time. Some time ago it was [https://docs.google.com/file/d/0B-droFdkDaVhVzhsN3RKRlFLODQ/edit?pli=1 shown] that cloud performance can be boosted by setting it on so naturally we decided to check this result. To make this test we issued requests for booting up and deleting VMs for different number of concurrent users ranging from one to 30 with and without this option set. For each group of users a total number of 200 requests was issued. Averaged time per request is shown below:
 
 
 
[[File:Amqp rpc single reply queue.png| amqp_rpc_single_replya_queue]]
 
 
 
So apparently this option affects cloud performance, but not in the way it was thought before.
 
 
 
 
 
 
 
== How To ==
 
 
 
<big>
 
# [[Rally/installation|Rally installation]]
 
# [[Rally/HowTo|How to use Rally]]
 
# [[Rally/DeployEngines|Available Deploy engines]]
 
# [[Rally/ServerProviders|Available Server providers]]
 
# [[Rally/BenchmarkScenarios|Available Benchmark scenarios]]
 
# [[Rally/HowToExtendRally|Extend Rally functionality]]
 
# [[Rally/RoadMap|Rally Road Map]]
 
</big>
 
 
 
 
 
 
 
== Weekly updates ==
 
Each week we write up on a special [[Rally/Updates|weekly updates page]] what sort of things have been accomplished in Rally during the past week and what are our plans for the next one.
 
 
 
 
 
== Join Rally team ==
 
 
 
==== Discussions & RoadMap ====
 
https://etherpad.openstack.org/p/Rally_Main
 
 
 
==== Open and assigned tasks ====
 
https://trello.com/b/DoD8aeZy/rally
 
 
 
To get account ping Boris in IRC (boris-42) or email me (boris(at)pavlovic.me)
 
 
 
==== IRC chat ====
 
server:  '''freenode.net'''
 
 
 
chanel: '''#openstack-rally'''
 
 
 
==== Weekly Meetings ====
 
The Rally project team holds weekly meetings on Tuesdays at [http://www.timeanddate.com/worldclock/fixedtime.html?msg=OpenStack+Rally+meeting&iso=20131029T21&p1=166&am=30 1700 UTC] in IRC, at the <code><nowiki>#openstack-meeting</nowiki></code> channel.
 
 
 
==== Source ====
 
https://github.com/stackforge/rally
 
 
 
==== Project space ====
 
http://launchpad.net/rally
 
 
 
==== Blueprints ====
 
active:    http://blueprints.launchpad.net/rally
 
 
 
v1 base: https://blueprints.launchpad.net/rally/+spec/init
 
 
 
==== Bugs ====
 
https://bugs.launchpad.net/rally
 
 
 
==== Pending Code Reviews ====
 
https://review.openstack.org/#/q/status:open+rally,n,z
 

Latest revision as of 12:59, 30 October 2017

The page has been moved to https://rally.readthedocs.io