Jump to: navigation, search

Difference between revisions of "Rally"

(Update for December 1, 2014)
(Documentation: fix broken link)
(22 intermediate revisions by 5 users not shown)
Line 3: Line 3:
 
== What is Rally? ==
 
== What is Rally? ==
  
If you are here, you are probably familiar with OpenStack and you also know that it's a really huge ecosystem of cooperative services. When something fails, performs slowly or doesn't scale, it's really hard to answer different questions on ''"what"'', ''"why"'' and ''"where"'' has happened. Another reason why you could be here is that you would like to build an OpenStack CI/CD system that will allow you to improve SLA, performance and stability of OpenStack continuously.
+
'''OpenStack''' is, undoubtedly, a really '''huge''' ecosystem of cooperative services. '''Rally''' is a '''benchmarking tool''' that answers the '''question''': ''“How does OpenStack work at scale?”''. To make this possible, Rally '''automates''' and '''unifies''' multi-node OpenStack deployment, cloud verification, benchmarking & profiling. Rally does it in a '''pluggable''' way, making it possible to check whether OpenStack is going to work well on, say, a 1k-servers installation under high load. Thus it can be used as a basic tool for an ''OpenStack CI/CD system'' that would continuously improve its SLA, performance and stability.
 
 
The OpenStack QA team mostly works on CI/CD that ensures that new patches don't break some specific single node installation of OpenStack. On the other hand it's clear that such CI/CD is only an indication and does not cover all cases (e.g. if a cloud works well on a single node installation it doesn't mean that it will continue to do so on a 1k servers installation under high load as well). Rally aims to fix this and help us to answer the question "How does OpenStack work at scale?". To make it possible, we are going to automate and unify all steps that are required for benchmarking OpenStack at scale: multi-node OS deployment, verification, benchmarking & profiling.
 
  
 
<center>[[File:Rally-Actions.png|850px]]</center>
 
<center>[[File:Rally-Actions.png|850px]]</center>
Line 13: Line 11:
 
* '''''Benchmark engine''''' - allows to create parameterized load on the cloud based on a big repository of benchmarks.
 
* '''''Benchmark engine''''' - allows to create parameterized load on the cloud based on a big repository of benchmarks.
  
For more information about how it works take a look at [[Rally#Architecture|Rally Architecture]]
 
  
 +
== Documentation ==
 +
'''[http://rally.readthedocs.org/en/latest/ Rally documentation on ReadTheDocs]''' is a perfect place to start learning about Rally. It provides you with an '''easy''' and '''illustrative''' guidance through this benchmarking tool. For example, check out the [https://rally.readthedocs.org/en/latest/quick_start/tutorial.html Rally step-by-step tutorial] that explains, in a series of lessons, how to explore the power of Rally in benchmarking your OpenStack clouds.
 +
 +
<center>[[File:Rally-ReadTheDocs.png]]</center>
  
 
== Use Cases ==
 
== Use Cases ==
Line 117: Line 118:
  
 
<big>
 
<big>
# [[Rally/installation|Install Rally]]
+
# [https://rally.readthedocs.org/en/latest/install.html Install Rally]
# [[Rally/HowTo|Use Rally]]
+
# [https://rally.readthedocs.org/en/latest/tutorial.html Rally step by step guide]
 
# [[Rally/RallyGates|Add rally performance jobs to your project]]
 
# [[Rally/RallyGates|Add rally performance jobs to your project]]
 
# [[Rally/Concepts|Main concepts of Rally]]
 
# [[Rally/Concepts|Main concepts of Rally]]
Line 126: Line 127:
 
:# [[Rally/Develop#How to contribute|How to contribute]]
 
:# [[Rally/Develop#How to contribute|How to contribute]]
 
</big>
 
</big>
 
== Updates ==
 
 
''Periodically, we write up on a special [[Rally/Updates|updates page]] what sort of things have been accomplished in Rally recently and what are our plans for the future. Below you can find the most recent report (December 1, 2014).'''
 
 
It's been a while since our last post here, and we've done quite a nice job in Rally during November. Let us share with you new things about Rally:
 
* '''Autogenerated HTML benchmark reports''' in Rally (which can be created by the '''''"rally task report"''''' command after a benchmark task has completed) have [https://review.openstack.org/#/c/131844/ been] [https://review.openstack.org/#/c/136435/ improved] further within the last month. As of now, the [http://logs.openstack.org/05/131005/28/check/gate-rally-dsvm-rally/f8f3da9/rally-plot/results.html.gz report page] contains an overview table, detailed informations about whether [http://logs.openstack.org/05/131005/28/check/gate-rally-dsvm-rally/f8f3da9/rally-plot/results.html.gz#/Authenticate.validate_cinder SLA (service-level agreement) checks] were successful and also detailed error logs, if any. Rally reports have become a wonderful tool indeed to analyse the benchmarking data as well as to share your results with others!
 
* Similar improvements have been made for HTML reports generated for the '''Tempest cloud verification''' ('''''"rally verify results --html --output_file <file>"'''''). New [https://review.openstack.org/#/c/135232/ enhanced] report pages have improved styling and refactored JS code.
 
* We have [https://review.openstack.org/#/c/137502/ changed] the way '''context classes''' in Rally should be declared. Having introduced a new '''''@context''''' decorator, we've made it much easier and also more readable.
 
* There is a new '''[https://review.openstack.org/127392 "servers" context]''' that allows you to create temporary servers before benchmark scenarios start and use these servers for testing inside these scenarios.
 
* '''New benchmark scenarios''' in Rally include those for '''[https://review.openstack.org/128631 Nova live migrate]''' and also a '''[https://review.openstack.org/127392 Cinder stress scenario]'''.
 
* '''Command-line interface improvements''' include an ability to [https://review.openstack.org/131463 refer deployments not only by uuid but also by name]. Please note that the syntax has changed a bit so now you have to supply the ''--deployment'' parameter to commands like ''"rally use deployment"'' (instead of ''--uuid'').
 
* There has been some '''major refactoring''' of the most critical parts of Rally code: the [https://review.openstack.org/129060 cleanup mechanism] and the [https://review.openstack.org/119297 "users" context code]. We are sure that after refactoring, this code has become both cleaner and less error-prone (as well as very pluggable in the case of cleanups).
 
 
 
Current work includes further code refactoring (e.g. in the ''Benchmark engine'' part), further CLI improvements (e.g. for the [https://review.openstack.org/131005 "rally task list" command]) and also new benchmark scenarios (e.g. for [https://review.openstack.org/137661 Murano]). We are also going to introduce a possibility of building Rally images for [https://review.openstack.org/#/c/132556/ Docker].
 
 
We encourage you to take a look at new patches in Rally pending for review and to help us make Rally better!
 
 
Source code for Rally is hosted at GitHub: https://github.com/stackforge/rally<br/>
 
You can track the overall progress in Rally via Stackalytics: http://stackalytics.com/?release=kilo&metric=commits&project_type=all&module=rally <br/>
 
Open reviews for Rally: https://review.openstack.org/#/q/status:open+rally,n,z
 
 
 
Stay tuned!
 
 
 
Regards,<br/>
 
The Rally team
 
 
 
'''[[Rally/Updates|Weekly Updates Archives]] '''
 
  
 
== Rally in the World ==
 
== Rally in the World ==
Line 164: Line 133:
 
|-
 
|-
 
! Date !! Authors !! Title !! Location
 
! Date !! Authors !! Title !! Location
 +
|-
 +
| 21/Jan/2015 || <ol><li> James Page </li></ol> || [https://javacruft.wordpress.com/2015/01/21/extreme-openstack-scale-testing-openstack-messaging Extreme OpenStack: Scale Testing OpenStack Messaging] || Ubuntu Server Team
 +
|-
 +
| 13/Jan/2015 || <ol><li> Vishal Yadav </li></ol> || [http://www.slideshare.net/vishalcdac/rally-openstackbenchmarking Rally: Testing & Benchmarking OpenStack] || India OpenStack Meetup Noida
 
|-
 
|-
 
| 29/May/2014 || <ol><li> Andrey Kurilin </li></ol> || [https://www.mirantis.com/blog/rally-openstack-tempest-testing-made-simpler Rally: OpenStack Tempest Testing Made Simple(r)] || https://www.mirantis.com/blog
 
| 29/May/2014 || <ol><li> Andrey Kurilin </li></ol> || [https://www.mirantis.com/blog/rally-openstack-tempest-testing-made-simpler Rally: OpenStack Tempest Testing Made Simple(r)] || https://www.mirantis.com/blog
Line 179: Line 152:
  
 
== Project Info ==
 
== Project Info ==
 +
 +
===Core team===
 +
 +
See [https://review.openstack.org/#/admin/groups/211,members current members]
 +
 +
=== Project meetings ===
 +
 +
See [[Meetings/Rally]]
  
 
=== Useful links ===  
 
=== Useful links ===  
* [https://github.com/stackforge/rally Source code]
+
* [https://github.com/openstack/rally Source code]
* [https://trello.com/b/DoD8aeZy/rally Rally trello board]
+
* [https://docs.google.com/a/mirantis.com/spreadsheets/d/16DXpfbqvlzMFaqaXAcJsBzzpowb_XpymaK2aFY2gA2g/edit#gid=0 Rally road map]
 
* [http://launchpad.net/rally Project space]
 
* [http://launchpad.net/rally Project space]
* [http://blueprints.launchpad.net/rally Blueprints]
 
 
* [https://bugs.launchpad.net/rally Bugs]
 
* [https://bugs.launchpad.net/rally Bugs]
 
* [https://review.openstack.org/#/q/status:open+rally,n,z Patches on review ]
 
* [https://review.openstack.org/#/q/status:open+rally,n,z Patches on review ]
* [http://eavesdrop.openstack.org/meetings/rally/2014/ Meeting logs]
+
* [http://eavesdrop.openstack.org/meetings/rally/2015/ Meeting logs]
 
* IRC [http://irclog.perlgeek.de/openstack-rally logs], server:  '''irc.freenode.net''' channel:  '''#openstack-rally'''
 
* IRC [http://irclog.perlgeek.de/openstack-rally logs], server:  '''irc.freenode.net''' channel:  '''#openstack-rally'''
 
'''NOTE:''' To be a member of trello board please write email to (boris at pavlovic.me) or ping in IRC boris-42
 
 
=== How to track project status? ===
 
 
The main directions of work in Rally are documented via '''''[https://blueprints.launchpad.net/rally blueprints]'''''. The most high-level ones are '''''*-base''''' blueprints, while more specific tasks are defined in derived blueprints (for an example of such a dependency tree, see the '''''[https://blueprints.launchpad.net/rally/+spec/benchmark-base base blueprint for Benchmarks]'''''). Each ''“base”'' blueprint description contains a link to a '''''google doc''''' with detailed informations about its contents.
 
 
While each blueprint has an assignee, single patchsets that implement it may be owned by different developers. We use a '''''[https://trello.com/b/DoD8aeZy/rally Trello board]''''' to track the distribution of tasks among developers. The tasks are structured there both by '''''labels''''' (corresponding to top-level blueprints) and by their '''''completion progress'''''. Please note the '''''Good for start''''' category containing very simple tasks, which can serve as a perfect introduction to Rally for newcomers.
 
  
 
=== Where can I discuss & propose changes? ===
 
=== Where can I discuss & propose changes? ===
  
 
# Our IRC channel: IRC server <code><nowiki>#openstack-rally</nowiki></code> on '''irc.freenode.net''';
 
# Our IRC channel: IRC server <code><nowiki>#openstack-rally</nowiki></code> on '''irc.freenode.net''';
# Weekly Rally team meeting: held 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 ('''irc.freenode.net''');
+
# Weekly Rally team meeting: held on Mondays at [http://www.worldclock.com/world_clock.html 1400 UTC] in IRC, in the <code><nowiki>#openstack-meeting</nowiki></code> channel ('''irc.freenode.net''');
 +
# Weekly release meeting: held on Mondays at [http://www.worldclock.com/world_clock.html 1300 UTC] in IRC, in the <code><nowiki>#openstack-rally</nowiki></code> channel ('''irc.freenode.net'''). At this meeting, we do not discuss common topics but release-specific stuff only, including the list of critical patches that have to be merged in the next release;
 
# Openstack mailing list: '''openstack-dev@lists.openstack.org''' (see [http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev subscription and usage] instructions);
 
# Openstack mailing list: '''openstack-dev@lists.openstack.org''' (see [http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev subscription and usage] instructions);
 
# [https://launchpad.net/rally Rally team on Launchpad]: Answers/Bugs/Blueprints.
 
# [https://launchpad.net/rally Rally team on Launchpad]: Answers/Bugs/Blueprints.

Revision as of 08:55, 3 February 2017


What is Rally?

OpenStack is, undoubtedly, a really huge ecosystem of cooperative services. Rally is a benchmarking tool that answers the question: “How does OpenStack work at scale?”. To make this possible, Rally automates and unifies multi-node OpenStack deployment, cloud verification, benchmarking & profiling. Rally does it in a pluggable way, making it possible to check whether OpenStack is going to work well on, say, a 1k-servers installation under high load. Thus it can be used as a basic tool for an OpenStack CI/CD system that would continuously improve its SLA, performance and stability.

Rally-Actions.png
  • Deploy engine is not yet another deployer of OpenStack, but just a pluggable mechanism that allows to unify & simplify work with different deployers like: DevStack, Fuel, Anvil on hardware/VMs that you have.
  • Verification - (work in progress) uses tempest to verify the functionality of a deployed OpenStack cloud. In future Rally will support other OS verifiers.
  • Benchmark engine - allows to create parameterized load on the cloud based on a big repository of benchmarks.


Documentation

Rally documentation on ReadTheDocs is a perfect place to start learning about Rally. It provides you with an easy and illustrative guidance through this benchmarking tool. For example, check out the Rally step-by-step tutorial that explains, in a series of lessons, how to explore the power of Rally in benchmarking your OpenStack clouds.

Rally-ReadTheDocs.png

Use Cases

Before diving deep in Rally architecture let's take a look at 3 major high level Rally Use Cases:

Rally-UseCases.png


Typical cases where Rally aims to help are:

  1. Automate measuring & profiling focused on how new code changes affect the OS performance;
  2. Using Rally profiler to detect scaling & performance issues;
  3. Investigate how different deployments affect the OS performance:
    • Find the set of suitable OpenStack deployment architectures;
    • Create deployment specifications for different loads (amount of controllers, swift nodes, etc.);
  4. Automate the search for hardware best suited for particular OpenStack cloud;
  5. Automate the 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.


Architecture

Usually OpenStack projects are as-a-Service, so Rally provides this approach and a CLI driven approach that does not require a daemon:

  1. Rally as-a-Service: Run rally as a set of daemons that present Web UI (work in progress) so 1 RaaS could be used by whole team.
  2. Rally as-an-App: Rally as a just lightweight CLI app (without any daemons), that makes it simple to develop & much more portable.


How is this possible? Take a look at diagram below:

Rally Architecture.png

So what is behind Rally?


Rally Components

Rally consists of 4 main components:

  1. Server Providers - provide servers (virtual servers), with ssh access, in one L3 network.
  2. Deploy Engines - deploy OpenStack cloud on servers that are presented by Server Providers
  3. Verification - component that runs tempest (or another pecific set of tests) against a deployed cloud, collects results & presents them in human readable form.
  4. Benchmark engine - allows to write parameterized benchmark scenarios & run them against the cloud.


But why does Rally need these components?
It becomes really clear if we try to imagine: how I will benchmark cloud at Scale, if ...

Rally QA.png


TO BE CONTINUED

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 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:

amqp_rpc_single_replya_queue

So apparently this option affects cloud performance, but not in the way it was thought before.


Performance of Nova instance list command

Context: 1 OpenStack user

Scenario: 1) boot VM from this user 2) list VM

Runner: Repeat 200 times.

As a result, on every next iteration user has more and more VMs and performance of VM list is degrading quite fast:

nova vm list performance

Complex scenarios & detailed information

For example NovaServers.snapshot contains a lot of "atomic" actions:

  1. boot VM
  2. snapshot VM
  3. delete VM
  4. boot VM from snapshot
  5. delete VM
  6. delete snapshot

Fortunately Rally collects information about duration of all these operation for every iteration.

As a result we are generating beautiful graphs:

snapshot detailed performance

How To

Actually there are only 3 steps that should be interesting for you:

  1. Install Rally
  2. Rally step by step guide
  3. Add rally performance jobs to your project
  4. Main concepts of Rally
  5. Improve Rally
  1. Main directions of work
  2. Where to begin
  3. How to contribute

Rally in the World

Date Authors Title Location
21/Jan/2015
  1. James Page
Extreme OpenStack: Scale Testing OpenStack Messaging Ubuntu Server Team
13/Jan/2015
  1. Vishal Yadav
Rally: Testing & Benchmarking OpenStack India OpenStack Meetup Noida
29/May/2014
  1. Andrey Kurilin
Rally: OpenStack Tempest Testing Made Simple(r) https://www.mirantis.com/blog
01/May/2014
  1. Boden Russell
KVM and Docker LXC Benchmarking with OpenStack http://bodenr.blogspot.ru/
01/Mar/2014
  1. Bangalore C.B. Ananth (cbpadman at cisco.com)
  2. Rahul Upadhyaya (rahuupad at cisco.com)
Benchmark as a Service OpenStack-Rally OpenStack Meetup Bangalore
28/Feb/2014
  1. Peeyush Gupta
Benchmarking OpenStack With Rally http://www.thegeekyway.com/
26/Feb/2014
  1. Oleg Gelbukh
Benchmarking OpenStack at megascale: How we tested Mirantis OpenStack at SoftLayer http://www.mirantis.com/blog/
07/Nov/2013
  1. Boris Pavlovic
Benchmark OpenStack at Scale Openstack summit Hong Kong

Project Info

Core team

See current members

Project meetings

See Meetings/Rally

Useful links

Where can I discuss & propose changes?

  1. Our IRC channel: IRC server #openstack-rally on irc.freenode.net;
  2. Weekly Rally team meeting: held on Mondays at 1400 UTC in IRC, in the #openstack-meeting channel (irc.freenode.net);
  3. Weekly release meeting: held on Mondays at 1300 UTC in IRC, in the #openstack-rally channel (irc.freenode.net). At this meeting, we do not discuss common topics but release-specific stuff only, including the list of critical patches that have to be merged in the next release;
  4. Openstack mailing list: openstack-dev@lists.openstack.org (see subscription and usage instructions);
  5. Rally team on Launchpad: Answers/Bugs/Blueprints.