2.1. Ocata OpenStack summit recap

During OpenStack Ocata Summit at Barcelona our team had several sessions and cross-teams discussions. This page contains links to the resources and related sessions recordings, that might be interesting for Performance Team members and newcomers.

2.1.1. Chasing 1000 nodes scale

This session is dedicated to our 1000 nodes emulation experiment we have done during Newton development cycle time frame. This test case aims to gain information about what resources and configuration are necessary for OpenStack control plane on large enough scale.

Useful links:

2.1.2. What Has Been Done During Newton Cycle and Ocata Planning

Results sharing / future planning session. All results and plans are logged to the Newton results / Ocata planning etherpad. Plans for Ocata development timeframe includes Mirantis, Inria, IBM and Red Hat participation.

2.1.3. OpenStack Scale and Performance Testing with Browbeat

Scalability and performance testing toolkit presentation by Red Hat. In this talk core Browbeat developers share how you can leverage Browbeat tool for your OpenStack deployment along with a demonstration showcasing some of the useful features. You can go through this presentation with Browbeat session summit video.

2.1.4. Massively Distributed Clouds Working Group

This is recently created WG dedicated to solve questions regarding the distributed OpenStack clouds - including questions about what performance and scalability the operator could expect from such deployments. All information about this group can be found using the following links:

2.1.5. Large Deployments Team meetup

During LDT working group summit meetup we collected a feedback from operators (listed in the LDT session etherpad).

2.1.6. Architecture Working Group

Architecture WG is a new group dedicated to be recognized forum of expertise on the design and architecture of OpenStack and to provide guidance and resources to the Technical Committee and the entire OpenStack community on architectural matters. This group is going to use similar methodologies with Performance Team help for architectural concepts verification and testing, as well as results presentation.

Useful links:

2.1.7. NFV Working Group

The WG follows on the previous telecom/NFV WG that met their goal of creating telecom/NFV related user stories. This new team will work with the OpenStack community to guide OpenStack Operators in running telecommunication services and Network Function Virtualization (NFV) systems. This means all kinds of things, from developing operational tools, to shepherding requirements through the OpenStack ecosystem and to performance testing and benchmarking. All information about this group can be found using the following links:

2.1.8. QA: destructive testing

During the design summit there was a separated QA session dedicated to destructive testing, and tool developed originally under Performance Team umbrella (os-faults library) was assumed to be #1 solution for those types of tests.

2.1.9. Cross Project Workshops: Rolling Upgrades, and the Road to Zero-Downtime

During this cross-project session Performance Team’s testing approach for reliability testing (when we’re injecting the destructive hooks) was mentioned as one of the opportunities for upgrades testing (when we’re injecting the upgrades process in the middle of testing and calculate exactly the same metrics we collected for reliability testing - mean time to recover, service downtime, performance degradation, etc.): upgrades summit etherpad.