Visit the OpenStack Summit page for the latest news, registration and hotels.

Sign up or log in to bookmark your favorites and sync them to your phone or calendar.

Operations [clear filter]
Tuesday, November 4


MySQL and OpenStack Deep Dive
Like it or leave it, there's no disputing that MySQL is by far the most commonly used database in OpenStack deployments. Most folks know that there are a number of different flavors of MySQL -- standard MySQL, MariaDb, Percona XtraDb, MySQL Galera, Percona XtraDb Cluster, etc. However, there are a number of misconceptions and myths around the pros and cons of these flavors of MySQL.

In this talk, we hope to dispel some of these myths and give some clear information on what are the strengths and weaknesses of each of these flavors, what pitfalls to avoid when migrating between them, and what architectural information to take into account when planning your OpenStack MySQL database deployments. We will discuss replication topologies and techniques, and explain how the Galera Cluster variants differ from standard MySQL replication.

Finally, in the latter part of the session, we do a deep dive into MySQL database performance analysis. We will be analyzing the results of a Rally run showing a typical Nova workload, and demonstrate using pt-query-digest to determine if a synchronously-replication database cluster like Percona XtraDb Cluster is a good fit for certain OpenStack projects.


Peter Boros

Principal Architect, Percona
Peter Boros joined Percona s European consulting team in May 2012. In the years before joining Percona he worked as a systems engineer and DBA at prominent companies including Sun Microsystems, where he specialized in performance tuning and was also a DBA at Hungary's largest social... Read More →
avatar for Jay Pipes

Jay Pipes

Amazon Web Services
I'm new to the Kubernetes scene but not open source. Please introduce yourself and say hi! Topics I'm interested in: reducing technical debt, documentation, scheduling, resource management and the Kubernetes contributor experience.

Tuesday November 4, 2014 11:15 - 11:55
Room 252 AB


Capacity Management/Provisioning (Cloud's full, Can't build here)
As a service provider, Rackspace is constantly bringing new OpenStack capacity online. In this session, we will detail a myriad of challenges around adding new compute capacity. These include: planning, automation, organizational, quality assurance, monitoring, security, networking, integration, and more. 

avatar for Andy Hill

Andy Hill

Systems Engineer, Rackspace, Rackspace Hosting
avatar for Joel Preas

Joel Preas

Public Cloud Systems Engineer
A cog in the machine at the Rackspace public cloud.
avatar for Matt Van Winkle

Matt Van Winkle

Senior Manager, Operations, Rackspace
I serve an amazingly talented group responsible for running the Rackspace public cloud. I have been with the company in various roles for over 13 years.  My primary task is helping my teams of engineers and developers build the software that manages our constantly growing fleet of... Read More →

Tuesday November 4, 2014 12:05 - 12:45
Room 252 AB


CI/CD Pipeline to Deploy and Maintain an OpenStack IaaS Cloud
In our role administering an OpenStack IaaS Cloud, we ve developed a release train allowing local development and testing of configuration management, testing in virtualised environments and automated deployment to staging and production, for the building and maintenance of an IaaS Cloud using upstream vendor OpenStack packages.

We will discuss the high-level concepts then review the details of our implementation and the tools we use, and created, to enable this pipeline.

We encourage software programming discipline to the administration of the Cloud infrastructure, with peer review, source code management and thorough testing before packaged releases.

Similarly, we encourage system administration discipline to configuration management code development, with deployments from scratch to volatile environments and upgrade deployments to stable environments.


Key elements of our approach are the use of upstream vendor packages for OpenStack (Ubuntu); configuration management (SaltStack); unit testing of configuration management (Test Kitchen with Kitchen-salt); Git, Gerrit and Gitshelf for source control management, peer review and packaging from git repos; volatile local development and remote test environments, using Vagrant and salt-cloud; and automated testing, packaging and deployment with the aid of Jenkins.


We will outline the CI/CD pipeline:

  1. Provision a multi-node, multi-network OpenStack development environment using Vagrant with Virtualbox (nova-qemu) with virtualised nodes representing a minimal region, and a salt-master with a file_roots tree built as per production.

  2. Develop salt configuration management code within this development environment. All installation and administration is managed with SaltStack, with both incremental upgrade and full bootstrap deployment testing.

  3. Push changes to Gerrit for both peer review and testing in discrete development environment instances, and unit testing with Test Kitchen via Jenkins.

  4. Automatically package a new release in a specific deploy-kit repo, using Gitshelf to build release tarball artifacts from a YAML file of code repo SHA1s.

  5. Automatically deploy to test environments in Public Cloud.

  6. Automatically deploy to real hardware in staging and production.


avatar for Simon McCartney

Simon McCartney

Technical Lead, HPE
Simon works for HP Cloud Services, having worked in both development & operations in companies large and small of the last 20 years in IT. Simon's current role in HP Cloud Services is as the Technical Lead for a private OpenStack cloud used to host public services. He lives in Northern... Read More →

Tuesday November 4, 2014 14:00 - 14:40
Room 252 AB


CI/CD in Practice: Real World Deployment and Management

What does it take to manage continuous automated delivery in 20 and counting openstack installations across the country?  What types of checks and balances do you have to make, how are upgrades most effectively implemented without failures?  How can we make the complexity of openstack simpler and more direct and ensure that we have a re-useable and scalable internal build and deployment process for all of these installs and upgrades?  

In this session, we will talk though how to design a CI/CD process that fits your organization, what has worked and what has blown up in our face along the way leading to the system that we have in place today.  From developer/ops best practices to custom packaging, repo management, and deployment we will be walking through a full lifecycle along with community tools available to save others time.  Without a consistant way to deploy and potentially patch and build custom openstack within your organization quickly, there is a risk of falling behind.  By designing a valid and well thought out CI/CD process and implementation this can be prevented!

avatar for sridhar basam

sridhar basam

Principal Engineer II, Comcast
Sri is the Lead Engineer at Comcast's Openstack based private cloud. In addition to Openstack, Sri's interests include performance optimization, the kernel, networking.
avatar for Prashanth Hari

Prashanth Hari

Senior Engineer, Comcast
Prashanth Hari is a senior engineer in Comcast Cloud. He is involved in Cloud automations and devops, building cloud performance metrics and monitoring tools, CI/CD, Tempest and Telemetry/Showback.
avatar for Victor Howard

Victor Howard

Principal Engineer, Comcast
Victor Howard is a Principal Software Engineer at Comcast and has been involved in CI/CD, Packaging, Keystone, and Heat.

Tuesday November 4, 2014 14:50 - 15:30
Room 252 AB


OpenStack Ops Using Serf and Puppet
While many OpenStack operators use Puppet to manage their deployment, either using home grown modules or via a product such as Fuel, orchestrating and reacting to changes in the cluster using Puppet can be difficult as Puppet has no elegant way to signal remote nodes. Serf is a relatively new project written in Go and using the Raft protocol that can be used to detect events within a cluster and react to those events.

As an Openstack operator, using the two in concert leads to a neat way to react to changes in cluster membership, load and failures using Serf without compromising the configuration management of the system via Puppet. This is done by running Puppet in a fully data driven way, with all data ingested from heira, and letting a Serf agent modify that data based on events it sees in the cluster before pushing events out to the rest of the cluster to realise the change.

This session will cover the basics of how we use Serf in conjunction with Puppet to manage production OpenStack clusters at Aptira, show off the potential benefits, some hiccups we discovered along the way and provide a basis for discussion of using cluster tools Serf with configuration management systems like Puppet.


Michael Chapman

Senior Software Engineer, Aptira
I've been running and hacking on OpenStack clusters in both testing and production environments since the Cactus release. I got my first taste running a relatively small cluster at a research institute in Australia before joining the team at Cisco to work on their OpenStack deployment... Read More →

Tuesday November 4, 2014 15:40 - 16:20
Room 252 AB


Performance Tuning a Cloud Application - A Real World Case Study

During the OpenStack summit in Atlanta, Symantec presented on our vision for a Key Value as a Service storage technology utilizing MagnetoDB.  Since then our Cloud Platform Team has rolled the service out in our production environments.  Through that process we have learned about tuning requirements of the solution on bare metal versus hosted VMs within an OpenStack environment.

Our initial performance testing was done with MagnetoDB running on bare metal nodes.  After migrating the service from bare metal to an OpenStack VM hosted environment, we observed a 50% reduction in performance.

This presentation will dig in to the technical details of performance tuning an application running on an OpenStack Nova Compute cluster.  We will examine the performance related configuration settings necessary to improve the hosted application from three different angles:

  1. the underlying compute node Operating System configuration

  2. the hypervisor virtualization layer

  3. and the Guest VM and Application stack 

We will provide an outline for other cloud operators to follow when they work towards performance tuning an application running in your own OpenStack cloud.  Specific topics that will be covered include:

  • compute hardware configuration and the impact on performance

  • compute operating system configuration and tuning

  • KVM hypervisor performance tuning parameters

  • guest VM configuration and how they relate to your application stack

  • guest Operating System configuration and tuning details

  • application performance tuning

At the end of this presentation you should have a guideline on how to approach performance tuning an application running in your OpenStack environment.

avatar for Shane Gibson

Shane Gibson

Sr. Principal Infrastructure Architect, Symantec, Corp.
Shane Gibson is the Sr. Principal Infrastructure Architect at Symantec for the Cloud Platform Engineering (CPE) team. He is responsible for the infrastructure design of the underlying platforms, operating systems, tools, and application stack that enables the OpenStack clusters within... Read More →

Tuesday November 4, 2014 16:40 - 17:20
Room 252 AB


Tuning OpenStack for Availability and Performance in Large Production Deployments

Symantec is building a large scale OpenStack based cloud platform to host products and services that provide security to our customers. As we build development, staging and production environments, we continue to learn and optimize our architecture and configuration for high availability, performance and security.


OpenStack has many moving parts that need to be taken into consideration when control services are deployed. We will share our learnings and deployment patterns of highly available and secure control services as well as challenges and issues that we are working on.  In this session we will share our learnings and will cover following topics:


  • RabbitMQ A/A clustering

  • Mysql A/A Clustering

  • Nova compute/api/conductor HA and tuning

  • Keystone

  • KVM

  • Config Management

Attendees will leave the session better prepared to architect their highly scalable, available and secure OpenStack control and data plane.


avatar for Gabriel Capisizu

Gabriel Capisizu

Cloud Infra Eng, Symantec
Gabriel Capisizu is part of the Symantec’s Cloud Platform Engineering team. He has over 10 years years of experience with large scale distributed systems, Unix and networking.  Gabriel has started with OpenStack in 2011 as part of PayPal’s cloud engineering. His focus is deployment... Read More →

Raj Geda

Cloud Eng, Symantec
Cloud Eng, Sys Admin

Tuesday November 4, 2014 17:30 - 18:10
Room 252 AB
Wednesday, November 5


Guest Image Lifecycle
There's lots of documentation on how to setup your openstack cloud, and lots of resources on what to offer tenants.

One aspect that seems underrepresented is the catalogue of guest images offered to users. 

Beyond just adding pre-baked images (like Ubuntu UEC), how does an openstack operator deal with maintaining a sane glance library, providing periodic updates to both packages and policies, Ideas for automation around deprecation and promotion of images.

Comparison of automated image creation tools, like packer, disk-image-builder.

avatar for Abel Lopez

Abel Lopez

Abe (a.k.a alop, the Happy bald guy) has been walking the fence between devops, engineering, deployer, and customer on openstack since 2012. Having previously been at AT&T doing large scale automation of openstack with chef, presently working on operationalizing it for Cisco.

Wednesday November 5, 2014 09:00 - 09:40
Room 252 AB


Learning to Scale OpenStack: Juno Update from the Rackspace Public Cloud

Rackspace invites the OpenStack community to hear the story of deployments at scale during the Juno cycle and share in the latest challenges, triumphs, and lessons learned from regularly upgrading the control plane of a production OpenStack public cloud. At the end of Icehouse, there were about 40,000 things in the public cloud to deploy to. How many deploy things will there be at the end of Juno? Did control plane releases speed up or slow down? Is Rackspace still looking at releasing every two weeks from upstream trunk?

Rainya Mosher will take you through the release and deploy process utilized within Rackspace, starting with a developer working on code upstream and ending with final delivery to six production environments. She ll highlight how changes to the packaging, gating, and deployment mechanisms have improved cycle times within the release and deploy process. She'll share stats on the state of OpenStack deployments at Rackspace, the impact the migration to Neutron had on release cycles, and performance improvements achieved in terms of speed and reduced customer impact. Rainya will end by looking forward to what is planned for the Kilo cycle that will make releases more visible and deployments less impactful for Rackspace and the entire deployer community.


Rainya Mosher

Product Manager, Digital, Rackspace Hosting
Rainya Mosher is a Product Manager for the Digital Practice Area at Rackspace Hosting . Prior to her transition into product, Rainya was the Engineering Dev Manager for the build release and deploy system for Rackspace OpenStack public cloud control plane. She has worked with a dozen... Read More →

Wednesday November 5, 2014 09:50 - 10:30
Room 252 AB


Monitoring OpenStack in the Enterprise
This talk will cover practical considerations for monitoring Openstack in the enterprise.  Jonathan Kelly, Openstack Architect and former Monitoring Architect at Rackspace, and Jake Briggs, Openstack Engineer at Rackspace will share their experience and practical guidance for monitoring production Openstack clouds.  The approach is product agnostic, so is applicable to any monitoring solution.

The audience can expect to walk away having learned:

* The benefits of a robust monitoring solution
* The difference between alerting and performance metrics, and why this distinction is critical
* The importance of a correlation and suppression strategy in removing noise from the monitoring system
* How to create an effective alerting system
* How to leverage alerting and performance metrics for service level validation
* How to leverage performance metrics for growth forecasting


Jonathan Kelly

Openstack Architect, Rackspace
Jonathan Kelly is an Openstack Private Cloud Architect at Rackspace. He has been involved with Openstack since the Bexar summit.

Wednesday November 5, 2014 11:00 - 11:40
Amphitheatre Bleu


Rethinking Ceilometer metric storage with Gnocchi: Time-series as a Service

When we started Ceilometer, we designed our metrics storage in a way that looked more like audit system that traditional metering tool.

Recent evolution in OpenStack and Ceilometer pushed us to rethink how we should treat and store metrics. This led to the creation of the Gnocchi project, a service providing time-series indexing and storage via an API, following the OpenStack philosophy.

When we started Ceilometer, we designed our metrics storage in a way that looked more like audit system than a traditional metering tool.

Recent evolution in OpenStack and Ceilometer pushed us to rethink how we should treat and store metrics. This led to the creation of the Gnocchi project, a service providing time-series indexing and storage via an API, following the OpenStack philosophy.

This talk will explain the design ideas underpinning Gnocchi, how it differs from the classic metering store provided by Ceilometer, and how it has been integrated into the Ceilometer metrics pipeline. We will also discuss the potential for performance & scalability improvements, and how the new metrics store will co-exist with the existing v2 Ceilometer API.

avatar for Dina Belova

Dina Belova

Senior Software Engineer, Mirantis, Mirantis
Dina has been working with OpenStack as a cloud platform for more than two years, creating private OpenStack-based clouds for a lot of customers, such as Gap, Dell and many others. Her experience includes close integration with all OpenStack projects and their improvement to satisfy... Read More →
avatar for Julien Danjou

Julien Danjou

Principal Software Engineer, Red Hat
I'm a Free Software hacker working at Red Hat. I've been contributing to open source projects for more than 15 years now, in various area, such as Debian, Freedesktop, awesome, Python or Emacs. Since 2011, I work on OpenStack Telemetry and act as its current Project Team Leade... Read More →
avatar for Eoghan Glynn

Eoghan Glynn

Principal Engineer, Red Hat
Eoghan is a Principal Software Engineer at the Red Hat OpenStack Infrastructure group, and is serving as Technical Lead for the OpenStack Telemetry Program over the Juno & Kilo cycles. Prior to OpenStack, Eoghan was at Amazon working on AWS monitoring services,.

Wednesday November 5, 2014 11:00 - 11:40
Room 252 AB


OpenStack Upgrade Without Down Time
In the openstack production environment, the service down time is sensitive to the customers. The system upgrade will be a very tough work for the operation team. 

In this session we will share the best practices and lesson learns about how to upgrade the openstack production environment, which was tested in the test environment with the same architecture and deployment of production environment. At the same time, we found some gaps and issues during the test which we want to address the attentions from the community to improve them.

The test was conducted for the upgrade from Havana to Icehouse. If time permit, another round of testing of upgrade from Icehouse to Juno will be performed before Paris Summit. If so we will share the test result and findings as well.

The topics to be covered in this session are:

1. The openstack production deployment architecture and design

2. The upgrade strategy and methodologies as well as the test methodologies.

3. The upgrade steps introductions

4. Best practicies during the upgrade

5. Lesson learns during the upgrade

6. Some known issues and gaps

(Canonical, NTT)


Yankai LIU

Cloud Architect, Canonical
Yankai Liu is a Cloud Architect at Canonical being responsible for cloud architecture design and delivery. She has more than 11 years IT industry experience in which with more than 4 years focusing on Cloud solution architecture and design. Before joining Canonical Yankai Liu worked... Read More →
avatar for Takashi Natsume

Takashi Natsume

Software Engineer, NTT, Nippon Telegraph and Telephone Corporation
Takashi Natsume has been working in the IT industry for about 15 years and has been working for NTT corporation since April, 2013.He is engaged in system design of public cloud systems based on OpenStack and functional verification of OpenStack.Prior to joining NTT corporation, He... Read More →

Wednesday November 5, 2014 11:50 - 12:30
Amphitheatre Bleu


The Road to Minimally Impacting Live Upgrades of the Rackspace Public Cloud
This is the story of Ada, an intrepid computer scientist, who has a need to create compute instances in the Rackspace Public Cloud. Alas, the public cloud is being upgraded at the moment and Ada's attempts to create compute instances are thwarted!

Upgrades of OpenStack can be difficult. Performing live upgrades on production clouds is even more arduous. Performing them live without the customer noticing is the hardest of all! In this talk, plucky Rackspace engineers Jesse Keating and Chris Blumentritt will recount the story of a Ada, our mythical user who tries desperately to make use of the Public Cloud during a control plane upgrade. They will provide narration, visualization, and technical insight into Ada's experience before and after making engineering and orchestration improvements to the control plane deployments during the Juno cycle. They will share the details of how they used Ansible playbooks, their Linux know-how, and  functionality available in Icehouse to provide less impactful live upgrades of OpenStack services in the Rackspace Public Cloud. 

Join us as we dive into the trails and tribulations and discover ways to prevent your "Adas" and other customers from ever noticing live upgrades of an OpenStack cloud so that you too can perform live upgrades that are 99% invisible.


Chris Blumentritt

Linux Systems Engineer, Rackspace
Chris is an awesome dude who does awesome things with Linux and OpenStack!

Jesse Keating

OpenStack Engineer, Blue Box
Jesse Keating is has been a part of the Linux community for over 13 years, as a user, contributor, instructor, author, and evangelist. As a believer in Continuous Integration and Continuous Delivery, currently Jesse is in a DevOps role at Rackspace, working on the Public Cloud.

Wednesday November 5, 2014 11:50 - 12:30
Room 252 AB


Long Periods of Boredom Punctuated by Moments of Terror'- Upgrading Live Openstack Clusters
Bloomberg runs several independent HA clusters based on the chef-bcpc archetecture.


Upgrading a live HA openstack cluster requires careful planning and guts of steal. We will talk about the experiences of upgrading Bloomberg s HA BCPC Openstack clusters. We will discuss the work flows we designed and tools we use for for upgrading the clusters and keeping them up to date. We also discus some of the things exciting things that can go wrong and share some of our scare tissue.  


Caius Howcroft

Cloud Builder

Wednesday November 5, 2014 13:50 - 14:30
Amphitheatre Bleu


Upgrading in Place from Grizzly to Icehouse: A Cautionary Tale
For their most recent major release, Piston decided to change the base of their OpenStack distribution from Grizzly to Icehouse. What could possibly go wrong? Plenty, of course! Platform Engineering Manager Forrest Carpenter walks through the gory details of the things Piston encountered -- most of them not anticipated -- during a sweeping change to the underpinnings of their product. What were the unknown unknowns, when should they have known about them, and how did the Piston engineers tackle them? Which components caused the most headaches, and how those were eventually solved, may surprise you.

avatar for Forrest Carpenter

Forrest Carpenter

Engineering Manager, Cisco
Forrest Carpenter is a former Distributed Systems Engineer with Piston Cloud and is currently an Engineering Manager with Cisco. With over two decades of experience in the high tech industry, most recently with VMware and Citrix, Forrest has contributed to both early stage startups... Read More →

Wednesday November 5, 2014 13:50 - 14:30
Room 252 AB


Are We Done Yet? Testing OpenStack Deployments
After constructing your OpenStack cloud, it can be difficult to determine whether you've actually configured all the components correctly. OpenStack Rally and Tempest have been created to help run verification and benchmarking tests for you, but they themselves are difficult to configure and use. This session will explore creating an easy and repeatable process verification and benchmarking process for your OpenStack cloud. Drawing on experience from numerous customer installations, it will delve into the benefits and pitfalls of using specific tools and technologies to achieve your testing goals.

avatar for Ken Pepple

Ken Pepple


Wednesday November 5, 2014 14:40 - 15:20
Amphitheatre Bleu


How We Fought for OpenStack HA
Anyone who s ever tried to build an OpenStack cloud knows how difficult it is to bring together all the components and tune them for optimized performance, especially for Highly Available installations. That s why so many deployment automation tools have emerged, including Fuel, Crowbar, Packstack, etc. Although it might seem straightforward to make all the OpenStack services HA after you complete a simple multi-host installation phase, be careful, as there will be dragons: AMQP message brokers not handling failover correctly, the database being put under a risk of corruption or need of manual reassembling, and the whole deployment finally being interfered by the way OpenStack messaging and database code is working.

In this talk, we ll explain how we re overcoming these problems in Mirantis OpenStack by using techniques such as:

  • Patching the vanilla OpenStack code

  • Optimizing deployment parts

  • Modifying Puppet providers and other tooling, to make the whole OpenStack cluster work as true Highly Available software

After attending this presentation, you ll know how to create a working HA cluster with either Mirantis OpenStack or vanilla OpenStack, and perhaps learn other techniques to help make your own installation more robust.

avatar for Sergii Golovatiuk

Sergii Golovatiuk

Lead Deployment Engineer, Mirantis Inc.
Kargo developer. K8S geek. Openstack HA contributor.

Vladimir Kuklin

Principal Deployment Engineer, Mirantis
Vladimir Kuklin is a Principal Deployment Engineer and Fuel Library Tech Lead at Mirantis. He has been working on Project Fuel since its inception in mid-2012, responsible for the end-to-end deployment cycle, from basic provisioning of OpenStack cluster nodes, DB, AMQP and HA configuration... Read More →

Wednesday November 5, 2014 14:40 - 15:20
Room 252 AB


Building the RackStack: Packaging from Upstream OpenStack
OpenStack is a rapidly evovlving code base, with over 1,000 commits to Nova and Glance projects during the Juno cycle as of August 2014. Upgrading the control plane from OpenStack's master branch throughout the release's development cycle, rather than at the end of the cycle, and keeping the production cloud aligned with upstream provides a unique set of challenges (and is more than a little insane.) Rackspace has been doing just that, however, for over 2 years. One of the keys to making releases from OpenStack master possible is the way in which we handle code delivery using a combination of software, automation, and process. In this talk, Philip Schwartz and Kevin Mitchell will pop the hood on the packaging solution they developed for Rackspace. Once OpenStack-Infrastructure CI is completed and a patchset is merged, Kevin and Philip's package picks up in order to create the "RackStack" payload from upstream and internal respositories as well as cloud-speciifc configurations for use upgrading the public cloud control plane. 

Topics Covered 

  • Branching and Tagging Strategy

  • Approach to Internal Patches

  • Merging Upstream and Local Using "PLY"

  • Unit Tests during Packaging

  • Building the Project Packages 

  • Improvements to Speed & Performance


Kevin Mitchell

Software Developer
Nova Core Reviewer and Rackspace Software Developer

Philip Schwartz

Senior Software Developer, Rackspace
I am a Senior Software Developer for Rackspace Hosting on the Build Release and Deployments Team for the past year. Prior to Rackspace I worked in the Big Data field as a core developer for HPCCSystems.com.  I have been working with OpenStack since the Austin cycle in both development/test... Read More →

Wednesday November 5, 2014 15:30 - 16:10
Amphitheatre Bleu


OpenStack High Availability: Are We There Yet?
Navigating the maze of High Availability approaches for OpenStack an independent comparison.

OpenStack has come a long way since it had its first discussions around infrastructure high availability in early 2012. Since then, every single OpenStack vendor has come out with their own flavor and understanding of high availability.

But how do they compare? What can and can t you do with them? And which should you choose?

This presentation compares the high availability support we currently have in OpenStack products from Red Hat, SUSE, Canonical, Mirantis, Cisco, Piston and others, from an independent and vendor-neutral perspective. It discusses advantages and disadvantages, deployment options, and suitable and not-so-suitable use cases.

avatar for Florian Haas

Florian Haas

VP Education, City Network
Florian runs the Education business unit at City Network, and helps people learn to use, understand, and deploy complex technology. He has worked exclusively with open source software since about 2002, and has been heavily involved in OpenStack and Ceph since early 2012, and in Open... Read More →

Wednesday November 5, 2014 15:30 - 16:10
Room 252 AB


Scaling OpenStack Services: The Pre-TripleO Service Cloud
In the world of clouds, horizontal scale is king. Running your OpenStack control services (APIs, message buses, databases, etc.) on a service cloud is a great way to achieve horizontal scale in a cloud friendly manner. Unfortunately, while OpenStack on OpenStack (TripleO) and OpenStack Bare Metal Provisioning (Ironic) are promising, they aren't yet ready for general consumption. While they mature it can be a daunting task to horizontally scale this infrastructure.

In this talk we will explore current, practical strategies for deploying a successful service cloud. Both the drawbacks and benefits of these immediate deployment paths will be covered, drawing on 3+ years of experience in running OpenStack clouds at enterprise scale.

avatar for Kevin Bringard

Kevin Bringard

Just Some Guy, Cisco Systems Inc.
Active participant in the OpenStack community. Designer of clouds. Maker of rain.
avatar for Richard Maynard

Richard Maynard

Engineer, ObjectRocket

Wednesday November 5, 2014 16:30 - 17:10
Room 252 AB


Tales from the Trenches: the Good, Bad, and Ugly of OpenStack Operations

Historically, some have considered OpenStack difficult to deploy and operate. There are dozens of resources from consultants and training to distributions and tools available to help, but is the problem really solved? The fact is, operating an OpenStack cloud is not a trivial endeavour.


This talk will detail what it takes to actually operate an OpenStack cloud in production, from the perspective of cloud operators who each have more than three years of first-hand experience operating OpenStack clouds across dozens of deployments. John and Craig will share with the audience best practices for OpenStack operations and discuss what works out of the box as well as the areas where improvisation is often required.


In addition to this operational overview, they will share war stories of measured successes, miserable failures and what was done to solve problems on production infrastructure.

This talk is intended to provide insight for fellow operators as well as prospective users.


avatar for John Dewey

John Dewey

Technical Leader, Cisco Systems

Craig Tracey

Solutions Engineer, Heptio
For the last 20 years Craig has helped build the infrastructure that powers the Internet. In this time he has had the opportunity to develop everything from kernel device drivers, to massive-scale cloud storage services, and even a few distributed compute platforms. Now as a Software... Read More →

Wednesday November 5, 2014 17:20 - 18:00
Room 252 AB