Loading…
This event has ended. View the official site or create your own event → Check it out
This event has ended. Create your own
View analytic

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

Monday, April 16
 

9:00am

Welcome

Introduction session on how the Design Summit works and how to make the most of it.


Monday April 16, 2012 9:00am - 9:25am
Ballroom

9:30am

Dependency management
Each project has a list of dependencies it requires and the version of each required. As the number of projects and the number of distros which OpenStack is packaged for grows, we need to reconcile these dependency sets in order to help downstreams do their thing. This session will brainstorm around how to do the following: * Define a set of dependencies and their minimum versions * Show which dependencies are common to all projects and which are specific to certain projects * Ensure that all projects work with the minimum version required and all later versions * Allow projects to have features which require newer versions while still supporting older versions * Have an inclusive cross-project process for adding new dependencies and updating their minimum versions which takes into account which target distros include those dependency versions (Session lead is Mark McLoughlin)

Monday April 16, 2012 9:30am - 9:55am
Seacliff AB

9:30am

Show-n-tell API site
Describe how the api.openstack.org site works. (Session lead is Anne Gentle)

Monday April 16, 2012 9:30am - 9:55am
Seacliff D

9:30am

OpenVZ Linux Containers
The talk is about operating system virtualization technology known as OpenVZ. This is an effective way of partitioning a Linux machine into multiple isolated Linux containers. All containers are running on top of one single Linux kernel, which results in excellent density, performance and manageability. The talk gives an overall description of OpenVZ building blocks, such as namespaces, cgroups and various resource controllers. A few features, notably live migration and virtual swap, are described in greater details. Results of some performance measurements against VMware, Xen and KVM are given. Finally, we will provide a status update on merging bits and pieces of OpenVZ kernel to upstream Linux kernel, and share our plans for the future. (Session lead is Kir Kolyshkin)

Monday April 16, 2012 9:30am - 10:25am
Bayview B

9:30am

RBAC, again...
Starting with an overview of what is now available with the Policy mechanism, discuss initial and early API access to allow application of Policy against existing services (Session lead is Joseph Heck)

Monday April 16, 2012 9:30am - 10:25am
Seacliff C

9:30am

Breakout of volume into its own service
The nitty gritty of how to separate nova-volumes into its own service and assigning action items. A rough outline of topics: Management: 1) Picking a name: cinder? 2) Founding of the core team 3) PTL election process 4) API design and management Action Plan: 1) Creating python-xxxxxclient 2) Alternative implementation of nova/volume/api.py that talks to client 3) Moving code into separate repo 4) Participation in openstack-common 5) Migration plan from Essex (Session lead is Vishvananda Ishaya)

Monday April 16, 2012 9:30am - 10:25am
Marina

9:30am

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Monday April 16, 2012 9:30am - 10:25am
Golden Gate

10:00am

Common configuration patterns
This session will attempt to build some consensus around common patterns for: * cfg usage * use (or not) of a global config object * default values * sample config files * common command line and configuration options * separating PasteDeploy config from user config (Session lead is Mark McLoughlin)

Monday April 16, 2012 10:00am - 10:25am
Seacliff AB

10:00am

Folsom documentation priorities
With the many docs sites listed at http://etherpad.openstack.org/folsomdocsplanning I'd like to discuss priorities for the upcoming release. We focused on the API user docs for Essex, what is a good focus for the Folsom docs? (Session lead is Anne Gentle)

Monday April 16, 2012 10:00am - 10:25am
Seacliff D

10:30am

Coffee Break
Monday April 16, 2012 10:30am - 11:00am
2nd Floor

11:00am

Folsom release schedule
The Folsom release schedule, exploring the option of releasing every 5-8 weeks instead of every 6 months. (Session lead is Thierry Carrez)

Monday April 16, 2012 11:00am - 11:25am
Seacliff AB

11:00am

Ryu Network Operating System
Ryu is an open-sourced Network Operating System (NOS) licensed under Apache License v2. The project URL is http://www.osrg.net/ryu/ . Ryu aims to provide a logically centralized control and well defined API that make it easy for operators to create new network management and control applications. Currently, Ryu supports OpenFlow protocol to modify the behavior of network devices. Ryu plugin for OpenStack was merged into Quantum (virtual network service) Essex. You can create tens of thousands of isolated virtual networks without using VLAN. The project goal is to develop an OSS Network Operating System that has high quality enough for use in large production environment in code quality/functionality/usability, and to become the standard network management component of OpenStack. In this talk, we will explain the overall design and future plans of Ryu, and give some demonstrations. (Session lead is Kazutaka Morita)

Monday April 16, 2012 11:00am - 11:25am
Bayview B

11:00am

Shared filesystems in nova
WikiMedia currently has a custom solution to create glusterfs filesystems that are shared among instances. Among other things, it allows all users of a given project to maintain a single home-directory across all instances. We'd like to move this functionality into OpenStack, either as a part of Nova or as part of the soon-to-be-spun-off Nova Volume service. I'll present a file-system-neutral design and demonstrate an implementation using Gluster as the back-end. (Session lead is Andrew Bogott)

Monday April 16, 2012 11:00am - 11:25am
Marina

11:00am

Content Sharing and Documentation
Describe ways we can share content and enable rebranding. Notes here: http://wiki.openstack.org/Documentation/ContentSharing (Session lead is Anne Gentle)

Monday April 16, 2012 11:00am - 11:55am
Seacliff D

11:00am

Trust delegation and federation
This crosses boundaries on a lot of topics; the intent is to allow for multiple instances of keystone and/or multiple instances of service endpoints, sharing and/or delegating trust to relevant systems. (Session lead is Joseph Heck)

Monday April 16, 2012 11:00am - 11:55am
Seacliff C

11:00am

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Monday April 16, 2012 11:00am - 12:25pm
Golden Gate

11:30am

Dough: OpenStack Billing Project
Presentation of a billing project for OpenStack projects. github.com/lzyeval/dough The items that we plan to charge our customers are instance usage, floating ip usage, Swift storage usage and throughput, and external network output. We have devised a billing system where the admin can register products of their cloud and periodically charge the user. Properties of a Product are region, item, item_type, payment_type, measure_unit, and price etc. Admin can register its own regions, items(instance, ips, network-io, storage-io), item_types(m1.tiny, m1.large), payment_types(hourly, daily, monthly), and map these four properties to a price and measure_unit. For example, when a user creates a instance from Horizon, a client code informs the billing API server and the billing system subscribes a instance product to the tenant_id. Based on the payment_type the user chose, the billing system will periodically check the status of the product and charge the tenant_id for its use. When the user terminates the instance, Horizon will send a unsubscribe request to the API server and stop charging the tenant from that point on. The billing system consists of three components. 1) API Server, which receives subscribe/unsubscribe and usage query requests. 2) Farmer, which periodically checks all subscriptions and requests billing to the collector. 3) Collector, gets work from the farmer and uses python-novaclient, kanyun-client, swift-client to retrieve usage information of a product the tenant has subscribed and charges money to the accordingly. The database used is MySQL and all communication between components uses ZeroMQ. Currently in testing and all source code is on Github. (Session lead is luozhongyue)

Monday April 16, 2012 11:30am - 11:55am
Bayview B

11:30am

Baking area for features
As of today, a feature is either "in" (merged in master branch) or "out" (nowhere). There is no way for experimental features to get peer-reviewed, bake and be tested until they are release-quality. This lowers the stability and quality of OpenStack projects in general. This session will explore how we can more efficiently separate one (or multiple) repo/branch with experimental features from the release repo/branch. The obvious model is the Linux kernel model, where features (when considered ready for release) get proposed to the mainline release branch during regular merge windows. (Session lead is Thierry Carrez)

Monday April 16, 2012 11:30am - 12:25pm
Seacliff AB

11:30am

Boot From Volume
There have been a number of discussion around boot from volume, what it actually means, what it should mean and what use cases are important. This session should be an opportunity to make some decisions on definitions and use cases for the Folsom timeline. (Session lead is John Griffith)

Monday April 16, 2012 11:30am - 12:25pm
Marina

12:00pm

Light-docs: articles, fast fixes, and priorities
Of course we want your documentation submissions in any format you want to write in. Bring all docs, RST, ASCIIdoc, markdown. Let's talk about how to produce a unified site once the docs are in a repository through automatic Jenkins builds. (Session lead is Anne Gentle)

Monday April 16, 2012 12:00pm - 12:25pm
Seacliff D

12:00pm

Atlas-LB and Yahoo's loadbalancer config system
Currently Atlas-LB has adapters for certain kinds of s/w and h/w Load balancers like Zeus and Netscalar. The adapters however, directly configure the LBs. If Atlas LB has to work with more complex data center topologies and complex business rules of different organizations, which have their own workflows for provisioning Load balancers, then it needs to be able to talk to the external workflow provisioning systems. This way, it is easy to design interoperability between Atlas, and other workflow systems. This proposal adds the capability for Atlas-LB to support a workflow load balancer provisioning system at the backend to which it could delegate LB management requests. It accomplishes this by having a new message listener which invokes an asynchronous workflow adapter in Atlas-LB which acts as a proxy to a workflow system. The current listener invokes the adapter synchronously and marks the loadbalancer/Vip as Active or in Error depending on whether the adapter returns successfully or not. For a workflow based system, since the adapter can not complete its task in a short/known amount of time (due to manual steps in the workflow), the adapter has to be asynchronous and the listener should not mark the loadbalancer/Vip as active even if the workflow adapter returns successfully. The successful return of the adapter is taken to mean only the successful creation of a request in the workflow system. (Session lead is Kiran Lonikar)

Monday April 16, 2012 12:00pm - 12:25pm
Bayview B

12:00pm

Keystone secret storage
By storing secret keys inside Keystone, other services can use that key to encrypt per-user data. In this way a security breach (e.g. database leak) won't leak any sensitive information. Further, these secret keys can themselves be stored in an encrypted inside Keystone, such that a Keystone database leak won't leak the secure keys either. Right now I have my own version of Keystone that implements this; I'd like to discuss how it works, and see if we can get similar functionality into the official Keystone server. (Session lead is justinsb)

Monday April 16, 2012 12:00pm - 12:25pm
Seacliff C

12:30pm

Lunch
Monday April 16, 2012 12:30pm - 1:30pm
2nd Floor

1:30pm

Lightning talks
Monday April 16, 2012 1:30pm - 1:55pm
Ballroom

2:00pm

The Resurrection of Burrow Simple Queue
Once upon a time, we had a little project that was almost incubated. It was going to provide OpenStack with a common, RESTful API for driving simple queues. Unfortunately, it was abandoned and withered on the vine. Recently, with the growing enthusiasm for OpenStack by current AWS users, I've found renewed interest in a simple OpenStack queue project. There has also been interest on the mailing lists for a common RPC mechanism. So I'm bringing burrow back from the dead. At the time of submission, it supports a significant subset Amazon Simple Query Service API and a single simple sqlite3 backend. Development is ongoing, so I expect a both a simple OpenStack REST API, and a direct API that can be used as a nova-rpc backend by the summit. Also, since I don't want to write a queue service, Burrow backs to real queue servers and a RabbitMQ backend should be done by then. (Session lead is ChristopherMacGown)

Monday April 16, 2012 2:00pm - 2:25pm
Bayview B

2:00pm

Keystone code structure
Andy Smith and/or Joe Heck to provide an overview of the internal components and internal Keystone (python) API for supporting: * Identity * Catalog * Token * EC2 * Service Explaining the code structure and how to extend to add additional backends. (Session lead is Joseph Heck)

Monday April 16, 2012 2:00pm - 2:25pm
Seacliff C

2:00pm

Compute cells presentation
Short presentation on compute cells, including the implementation, current status, and ideas for cleaning it up. (Session lead is Chris Behrens)

Monday April 16, 2012 2:00pm - 2:25pm
Marina

2:00pm

openstack-common: progress and plans
This session will review the progress made during the Essex cycle towards making openstack-common a reality and the principles we've settled on for the project (http://wiki.openstack.org/CommonLibrary). We will discuss the next steps for the project and attempt to line some people up to help with specific goals for the Folsom cycle. It will include a discussion about moving RPC mechanism and notifications to common. The notification implementations have reportedly diverged in Nova and Glance and are dependent on a common RPC mechanism, although no code reuse currently exists today. (Session lead is Mark McLoughlin)

Monday April 16, 2012 2:00pm - 2:55pm
Seacliff AB

2:00pm

Discuss XenAPI (XenServer/XCP) Folsom Roadmap
Have a lot of the people working on XenAPI together in one room, and prioritise what we think is important for Folsom. Think about: * review differences with libvirt * update on live migration without the need for shared storage and pools (coming soon in XCP & XenServer) * discuss any issues blocking people using XenServer * create plans to expand automated QA, particularly around multi-server setups, going from baremetal to Hypervisor with OpenStack domU (Session lead is John Garbutt)

Monday April 16, 2012 2:00pm - 2:55pm
Seacliff D

2:00pm

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Monday April 16, 2012 2:00pm - 3:55pm
Golden Gate

2:30pm

How to allow/enable multifactor auth
How to allow for/enable multifactor authentication - pluggable backend for multiple authN sources (ex: mobile authN from verisign but SMS done through Telesign) - Potential out-of-box integration with WikiD - an opensrouce MFA provider- allowing MFA for different tenants/users. Ex: Access to tenant A requires 3 authN but tenant B requires 2. User Jane requires 3 authN but user test_service requires 1. (Session lead is Joseph Heck)

Monday April 16, 2012 2:30pm - 2:55pm
Seacliff C

2:30pm

Distributed RPC: code & architecture analysis
A ZeroMQ-based RPC plugin is now available for Essex and seeking inclusion in Folsom. This new driver differs from OpenStack's existing RPC implementations by implementing a distributed, peer-to-peer design. I will explain why this driver is important and how it works. (Session lead is Eric Windisch)

Monday April 16, 2012 2:30pm - 2:55pm
Marina

2:30pm

PyPI, Client Libraries, and CLP
Replacing "Installing OpenStack CI", we'll be brainstorming around our client lib releases and intereaction with PyPI. (Session lead is Monty Taylor)

Monday April 16, 2012 2:30pm - 3:25pm
Bayview B

3:00pm

Standardizing database management across projects
Each Openstack project currently has its own unique way of managing its migration repositories + sqlalchemy models and its own expectations in terms of upgrades. Worse, all gating of database code currently targets a single database and configuration (whatever devstack happens to be using). In the interest of providing users, packagers and distributors with a consistent and reliable upgrade experience across projects, we should standardize on some common practices and improve test coverage of database code. (Session lead is Adam Gandelman)

Monday April 16, 2012 3:00pm - 3:25pm
Seacliff AB

3:00pm

PKI support
Adam Young: Authentication in OpenStack is a two part mechanism. The first stage is when the user makes the initial authentication to Keystone, which results in the issue of a token. The second is the use of the token to provide single sign on and delegated authentication throughout the OpenStack cluster. PKI can improve the security of the first stage. It can both help security and scalability of the second. (Session lead is Joseph Heck)

Monday April 16, 2012 3:00pm - 3:25pm
Seacliff C

3:00pm

Bare metal cloud on commodity hardware
Intel's vPro offers remote power-on/power-off, boot redirection, network controls and video redirection. That's exactly the feature set we need for bare-metal cloud. Intel vPro is available on many computers, with a minimal price premium: if you are reading this on a computer bought by a large company, your machine likely has everything it needs to become a node in an OpenStack bare metal cloud. Having a bare-metal cloud with commodity hardware would: * simplify testing and deployment of OpenStack * allow cloud customers to get full, bare machines through the same API * unify the management of dedicated machines & cloud machines for providers * eliminate the virtualization tax (e.g. for databases) I've done a number of experiments with Intel's vPro, and I believe it does support all the functionality we need. We should also discuss whether we can achieve this through PXE, IPMI or other remote control systems. (Session lead is justinsb)

Monday April 16, 2012 3:00pm - 3:55pm
Seacliff D

3:00pm

Thread model and concurrency
We should revisit the use of eventlet (hence green threads) as the basis of the thread model. We should talk about it's pros and cons, and alternatives, such as fall back to vanilla Python threads, or to use multiprocessing. This should apply to nova, glance and keystone. The locking mechanisms, including "FOR UPDATE" clauses in db, file based locks and in memory mutex/semaphore should be considered in the thread model change. (Session lead is Yun Mao)

Monday April 16, 2012 3:00pm - 3:55pm
Marina

3:30pm

Stable branch maintenance and releases
During the Essex cycle, we established a policy for stable branches (http://wiki.openstack.org/StableBranch), maintained the stable/diablo branches for Nova and Glance and pushed some 2011.3.1 releases. This session will review our progress and make plans for the Folsom. Topics covered will include the stable branch policy, release process, the projects which we will maintain stable/essex branches for, tools we can use and appeal for new members to join the stable team. (Session lead is Mark McLoughlin)

Monday April 16, 2012 3:30pm - 3:55pm
Seacliff AB

3:30pm

Common Testing Interface
Over the last cycle we arrived on a common testing interface which jenkins uses to run the various tests on a given repo tree - but it was emergent rather than pre-planned. We should sit down, go through the requirements, make sure we like them, and see what we might want to add. Additionally, planning ahead for what a similar interface might look like for non-python projects (should they arise) would be a nice way to be prepared. (Session lead is Monty Taylor)

Monday April 16, 2012 3:30pm - 3:55pm
Bayview B

3:30pm

Middleware and additional information flow
Discussion of middleware - - where to house it - openstack common overlap - additional/optional metadata desired for passing (Session lead is Joseph Heck)

Monday April 16, 2012 3:30pm - 3:55pm
Seacliff C

4:00pm

Coffee Break
Monday April 16, 2012 4:00pm - 4:30pm
2nd Floor

4:30pm

Jenkins jclouds plugin
We've been working on the development of a plugin for jenkins that lets us automatically provision build and test nodes as needed - which allows us to take advantage of all of the clouds we have for both speed and resiliency in our build farm. Most of how that's happening is self-contained in the CI team - so letting people know how that's happening and how they can take advantage of it I thought would be helpful. (Session lead is Monty Taylor)

Monday April 16, 2012 4:30pm - 4:55pm
Bayview B

4:30pm

Performance and Caching in Nova
Things we're doing to make nova faster. Will talk about performance measurements with NovaProf, Tach (w/Statd & Graphite) and StackTach. (Session lead is Sandy Walsh)

Monday April 16, 2012 4:30pm - 4:55pm
Marina

4:30pm

OpenStack client tools (or unified tool)
Currently all of the OpenStack cli tools are named after the project codename and use varying styles of commands and options. While we made a bit of progress in this regard during Essex, a number of major changes need to be made to align all 5 clients. Since a name change is one of them maybe this is a good time to start clean. As a fallback, using a common command format, options format, etc across the projects would be a huge improvement. Some questions to address: - Preferences for command format, eg ' ' vs '-' - Consolidation of common client functions while still using the actual client modules. - Where would this work be housed in the OpenStack hierarchy? See http://wiki.openstack.org/UnifiedCLI for blueprint. (Session lead is Dean Troyer)

Monday April 16, 2012 4:30pm - 5:25pm
Seacliff AB

4:30pm

v.Next REST API
Workshop to nail down a prioritized list of additional/expanded REST API for Keystone, revising the existing API to cover holes found in implementations between Diablo and Essex, and to expand to allow future federation capabilities, etc - based on brainstorm sessions in earlier sessions. (Session lead is Joseph Heck)

Monday April 16, 2012 4:30pm - 5:25pm
Seacliff C

4:30pm

Hyper-V Revival
Demo of updated hyper-v driver working on the latest Essex code with a full installer. Discussion around the complexity of coding for Hyper-V on Windows via Windows or Linux and how to test with limited gear. Call for beta testers/companies interested in using Hyper-V and a call for developers interested in working on the code base for a full port to windows and re-inclusion in Folsom. (Session lead is Jordan Rinke)

Monday April 16, 2012 4:30pm - 5:25pm
Seacliff D

4:30pm

The Open Cloud computing Interface & OpenStack

Topics to be discussed include: What is OCCI and its goals? Where does OCCI fit the OpenStack picture? What does the the OpenStack OCCI API implementation do today? How should OpenStack address “extra” APIs? Q&As. All are welcome to it from the inquisitive to the skeptical!

Join Eugene Luster (R2AD).


Monday April 16, 2012 4:30pm - 5:25pm
Golden Gate

4:30pm

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Monday April 16, 2012 4:30pm - 5:55pm
Golden Gate

5:00pm

Openstack Notifications System and Yagi.
A brief overview of the Notifications system used by Nova, Glance and other Openstack projects, plus some more in-depth detail on integrating external systems, such as billing and monitoring systems, with Openstack using the Yagi feed-radiator application (https://github.com/Cerberus98/yagi/) to publish notifications in Atom format thru AtomPub or PubSubHubub protocols. (Session lead is Monsyne Dragon)

Monday April 16, 2012 5:00pm - 5:55pm
Bayview B

5:00pm

Scaling Openstack
There are several aspects of scaling openstack for very large deployments. Some include * scheduling (including network aware) * data store scalability * network layer scalability (state management) * scalable storage (constraint aware scheduling, caching) * scalable monitoring of large deployments (Session lead is Debo~ Dutta)

Monday April 16, 2012 5:00pm - 5:55pm
Marina

5:30pm

Python 3 and 'G'
This is a brainstorming session to determine what we can or should do during the Folsom release cycle to prepare for Python 3 in a G or later release. After PyCon 2012, it's clear than Python 3 is here, and we need to starting thinking about it. It's also clear this isn't happening in Folsom. This is a big effort, and we should probably start planning sooner rather than later. Initial topics for discussion are: - Tracking our package dependencies and their Python 3 status - Will all projects move to Python 3 simultaneously ? - How do we handle incubated projects and contributed modules ? - Do we even consider supporting 2.x and 3.x simultaneously ? - Impacts on continuous integration - Impacts on downstream packages and distros - What code guidelines can we put in place for Folsom to help the transition ? The objective of this session to produce a short list of specific actions we an take in the Folsom timeframe to make the inveitable transition easier. (Session lead is Mike Pittaro)

Monday April 16, 2012 5:30pm - 5:55pm
Seacliff AB

5:30pm

Token API
Guang Yee to proposed significant revisions to the /token REST API (Session lead is Joseph Heck)

Monday April 16, 2012 5:30pm - 5:55pm
Seacliff C

5:30pm

Image caching, where to from here
As of essex libvirt has an image cache to speed startup of new instances. This cache stores images direct from glance, as well as resized images. There is a periodic task which cleans up images in the cache which are no longer needed. Do we want to implement this for other hypervisors as well? Additionally, the blueprint (https://blueprints.launchpad.net/nova/+spec/nova-image-cache-management) proposed that popular / strategic images could be pre-cached on compute nodes. Is this something we still want to do? What factors do we want to use for the reference implementation? (Session lead is Michael Still)

Monday April 16, 2012 5:30pm - 5:55pm
Seacliff D

6:00pm

Mirantis Summit Kickoff Party

http://openstacksummitkickoff2012.eventbrite.com/

Mirantis, the company famous for bringing the finest in white night vodka to OpenStack events, invites you to our OpenStack Summit Kickoff Party, Monday night April 16th. Join us to celebrate the formal release of Essex, the kickoff of Folsom, the OpenStack Summit and Developer Conference, and April in San Francisco at the Sens San Francisco Restaurant, in Embarcadero center just next door to the Hyatt from 6-9 pm, while the night is still white. Take in ample food, drinks, some slightly off-the-wall entertainment and a good time getting cloudy enough to make the conference a little damned smashing.


Monday April 16, 2012 6:00pm - 9:00pm
Sens Restaurant Embarcadero Center, San Francisco

8:30pm

Art and OpenStack with CloudScaling, SolidFire and Rightscale

Please RSVP: http://artandopenstack.eventbrite.com/

Join Cloudscaling, SolidFire and RightScale at 111 Minna Gallery on the opening night of the OpenStack Design Summit. We'll celebrate the success of one of the fastest-growing projects in open source history.

The party starts at 8:30 pm on the first night of the Summit: Monday, April 16. Enjoy an awesome menu by Foxtail Catering, music by KZSU Stanford's Holt Sornsen, dancing, libations and a fabulous show by San Francisco artist Alec Huxley.

The Gallery is a 5-minute cab ride from the Hyatt, and we'll have shuttles running from the main lobby entrance of the hotel starting at 8:30 on 15-minute intervals.

And be sure to join our friends at Mirantis earlier in the evening. They're celebrating at Sens Restaurant next door to the Hyatt Embarcadero starting at 6:00 pm.


Monday April 16, 2012 8:30pm - 11:00pm
111 Minna Street Gallery 111 Minna st, San Francisco
 
Tuesday, April 17
 

9:00am

Chef and OpenStack
Chef is a configuration management and service automation tool frequently used in the OpenStack community. There have been a number of threads on the mailing list around its usage and the various implementations. It would be nice to get an overview of the landscape and help people coordinate efforts where possible. I will give a brief introductory presentation of the existing Chef cookbooks I'm aware of and the use of knife-openstack with Nova. After that we'll open it up for discussion. (Session lead is mattray)

Tuesday April 17, 2012 9:00am - 9:55am
Bayview B

9:00am

Splitting Nova and making it more pluggable
Nova is too monolithic. It's difficult to develop external components without affecting the core. Nova-core reviewers are expected to be specialists in everything. We need to split Nova into implicit (same project) or explicit parts (separate projects). How can this be accomplished ? What are the no-brainer candidates ? Where are the pain points ? What can we plan to do over the Folsom timeframe (Session lead is Thierry Carrez)

Tuesday April 17, 2012 9:00am - 9:55am
Seacliff AB

9:00am

Test Strategy, Processes, and Quality Metrics
Brainstorm session to cover: Test strategy,processes, and Quality Metrics. 1) Test Strategy: Test Strategies to improve quality for Folsom release. 2) Test process Definition: Discussion and concurrence on a) Types of tests – Terminology of smoke, unit, functional, integration, and performance testing. b)Bug Severity definition c) Test coverage Test coverage analysis 3) Quality – Key metrics Discussion with community to identify top 5 key metrics to assess release quality (Session lead is Ravikumar Venkatesan)

Tuesday April 17, 2012 9:00am - 9:55am
Seacliff C

9:00am

Quantum Overview for Developers
Last summit, we found that many people attended Quantum design sessions, but where not familiar with Quantum as a platform, and thus, weren't able to effectively contribute. This session will be an overview of the basics of Quantum, with a 20 minute presentation, and the rest of the time slotted for questions. The goal is to get people fairly new to Quantum up to speed quickly. This session is "required" for anyone who is looking to participate in Quantum design discussions, but is not already familiar with Quantum from a developer perspective. (Session lead is dan wendlandt)

Tuesday April 17, 2012 9:00am - 9:55am
Seacliff D

9:00am

Swift Quotas
In this session I would like to: * determine if there really is support for this feature * hammer out the requirements * see if I can interest any Swift developers in implementing this feature :) * possibly discuss the design and implementation * see if there's any crossover with the Quota Classes stuff See: * https://bugs.launchpad.net/swift/+bug/903350 * https://blueprints.launchpad.net/swift/+spec/storage-quotas * http://wiki.openstack.org/SwiftQuotas Also related: * https://lists.launchpad.net/openstack/msg08775.html * https://blueprints.launchpad.net/nova/+spec/quota-classes * http://wiki.openstack.org/QuotaClass (Session lead is Everett Toews)

Tuesday April 17, 2012 9:00am - 9:55am
Marina

9:00am

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Tuesday April 17, 2012 9:00am - 10:25am
Golden Gate

10:00am

Service relationships
A volume service may be associated with an explicit compute service. There is no way to explicitly indicate this relationship via the service catalog today. Type and names can be duplicated according to the contract and expanding a URL type for a relationship seems like over-kill. There needs to be a collection or grouping of services to indicate which ones play well with each other. (Session lead is Joe Savak)

Tuesday April 17, 2012 10:00am - 10:25am
Bayview B

10:00am

Versioning the Nova RPC APIs
Nova has an RPC mechanism used to communicate between services. There is no way to detect potential incompatibilities between services due to version skew. I would like to discuss approaches to make these interfaces more robust so that failures due to interfaces not matcing up can be explicit and obvious. (Session lead is Russell Bryant)

Tuesday April 17, 2012 10:00am - 10:25am
Seacliff AB

10:00am

Automating complex Openstack deployment testing
The Openstack Ubuntu Testing project has delivered automated testing of OpenStack Essex on Ubuntu Precise on actual hardware during the 12.04/Essex development cycle. 1) Quick overview of Openstack Ubuntu Testing and current activities. 2) Discussion: - How do we integrate this better into the OpenStack development process to make testing a two way process? - Pre-commit testing on merge proposals? - Feedback into gerrit? - Dealing with issues related to packaging, not openstack - Target hardware configurations - Network topologies - Target components - Currently core only - nova, keystone, glance - Expand to cover swift, melange, quatum, horizon? - Virtualization types - kvm and/or lxc - Juju charm management for testing - Reusing this framework for virtualized testing. (Session lead is James Page)

Tuesday April 17, 2012 10:00am - 10:25am
Seacliff C

10:00am

Melange Overview
We are going to be talking about how to merge at least portions of the Melange IPAM capabilities into Quantum. As a prelude to that discussion, we want to get a description of the current state of Melange. (Session lead is dan wendlandt)

Tuesday April 17, 2012 10:00am - 10:25am
Seacliff D

10:00am

Aligning Swift with the rest of OpenStack
Swift is often considered (for good and bad reasons) as the black sheep of the OpenStack flock. It's the only OpenStack core project using specific versioning, frequent releases, feature-based/unpredictable schedule, internal QA sign-off... Can we (should we) change things where possible to fix that perception ? Where can Swift align with the other projects ? Where should the other projects align with Swift ? Where should the difference just be understood and accepted ? (Session lead is Thierry Carrez)

Tuesday April 17, 2012 10:00am - 10:25am
Marina

10:30am

Coffee Break
Tuesday April 17, 2012 10:30am - 11:00am
2nd Floor

11:00am

Host Aggregates v2
As per recent discussion in the mailing list - discus how to expose to the scheduler which group of host support a particular abstract subset of flavors/images/capabilities Host Aggregates may provide the base for this - but it is currently tied to hypervisor tyep capabilities (Session lead is Phil Day)

Tuesday April 17, 2012 11:00am - 11:25am
Seacliff AB

11:00am

Integrating Openstack Swift with Cloudstack
A lot of buzz is around Cloudstack compute stack but there is currently a lack of integration with the openstack projects. We created a swift authentication middleware that authenticates through cloudstack. With the current Cloudstack integration that support swift for templates/snapshots backend, this middleware facilitate the rest of the integration needed to also use swift as an object store service with cloudstack. (Session lead is David Chamard)

Tuesday April 17, 2012 11:00am - 11:25am
Marina

11:00am

Performance Evaluation for OpenStack
OpenStack has been more popular than ever and more vendors in China today are trying build their own products/solutions based on OpenStack. To achieve product-level quality for Folsom, performance, scalability and power efficiency should be as important as feature readiness and robustness. We'd like to share some of our proof-of-concept performance work on Nova - and we really looking forward to discussion on general approach on how to evaluate and improve OpenStack solution performance (e.g. workload, benchmark methodology, related optimization effort etc) and maybe start some projects work on it. (Session lead is Huang Zhiteng)

Tuesday April 17, 2012 11:00am - 11:55am
Bayview B

11:00am

Smoke Testing Realistic Deployments
How can we improve our smoke testing pipeline so that we catch issues in merge proposal *before* they land? A brief overview of SmokeStack. What it is good at... and how the system might be improved. Discussion Questions: -Gating vs. Pre-gating (commenting on merge proposals)? -How to deal with false negatives? -Speed improvements? Get more resources? Better utilize resources? Does speed matter? -What configurations are the most important? (Libvirt, XenServer, MySQL, PostgreSQL, LDAP, etc) -Multi-node vs. Single node. -Using/building real packages? (Fedora/Ubuntu) -How does config management fit in? (Chef/Puppet) (Session lead is Dan Prince)

Tuesday April 17, 2012 11:00am - 11:55am
Seacliff C

11:00am

AuthN/AuthZ and RBAC for Quantum
There are a number of use cases driving the need for more fine grained controls/permissions within Quantum. This session will collect core use cases and discuss a roadmap that quickly solves some tactical Auth needs while laying the groundwork for a more comprehensive solution during the Folsom development timeframe. The goals will be a prioritized set of blueprints and an initial mapping of those blueprints against Folsom milestones. (Session lead is Troy Toman)

Tuesday April 17, 2012 11:00am - 11:55am
Seacliff D

11:00am

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Tuesday April 17, 2012 11:00am - 12:25pm
Golden Gate

11:30am

Implementing an IaaS Standard (CIMI) for Nova
The DMTF is working with a number of Cloud vendors and consumers to develop an IaaS API called CIMI (Cloud Infrastructure Management Interface). This sessions is focused on designing an implementation as an optional alternative to the existing OpenStack and EC2 APIs. A brief overview of the specification will be given with a mapping of CIMI concepts and APIs to existing OpenStack APIs. Discussions will be centered on where the implementation would exist, on the ideal architecture for invoking APIs across the entire spectrum of OpenStack projects, and on addressing incompatibilities between the spec and existing OpenStack APIs. (Session lead is Doug Davis)

Tuesday April 17, 2012 11:30am - 11:55am
Seacliff AB

11:30am

Enabling Swift Versioning and Distributed Dedup
The proposal is to add versioning support to Swift by adding distributed deduplication support at the same time. An alternate ring is proposed that breaks objects into uniquely named Version Manifests and deduplicated Chunks. The presentation will cover how this system would work and how it would co-exist with current Swift rings. An etherpad description of the project has been posted to http://etherpad.openstack.org/P9MMYSWE6U (Session lead is Caitlin Bestler)

Tuesday April 17, 2012 11:30am - 12:25pm
Marina

12:00pm

Fog and OpenStack Developer Meetup
Fog is the Ruby cloud services library (http://fog.io/) which provides integration with a number of cloud providers, including OpenStack. There are/have been a number of duplicated efforts working with Fog. I propose a developer session to help coordinate future efforts and put names with faces. (Session lead is mattray)

Tuesday April 17, 2012 12:00pm - 12:25pm
Bayview B

12:00pm

Improvements to Nova Service Management
This is a collection of related ideas to improve the management of Nova Services: i) Add a text field to capture the reason a service is disabled. Disabling a specific services is normally a manual operation, carried out for some specific operational reason, such as an issue on a specific server, or to reserve some servers as part of capacity management. We’ve found that adding a reason field which can be set when the service is disabled and displayed as part of the service status makes it much easier to manage this aspect of the system. The messages are normally short and refer to, for example, issues captured in JIRA. ii) In addition to updating the services table with a timestamp and report count, services should report the number of threads currently running in their greenpool – as this provides an indication of how busy the service is. This information could be used, for example, by schedulers. iii) Extend the services table to include a status field that is set by the service to “starting/started/stopping/stopped” at the appropriate points in its lifecycle. This helps distinguish between services that have stopped as part of a software update from those that have failed for some reason. iv) Service shutdown currently kills any threads which are in progress. Service shutdown should try to stop the service in a clean state by stopping reading new messages from the queue and waiting for the running thread count to reduce to 0. If the threads don’t stop within a configurable timeout window, then the service should force a stop anyway. (Session lead is Phil Day)

Tuesday April 17, 2012 12:00pm - 12:25pm
Seacliff AB

12:00pm

Performance Testing OpenStack
1. Demonstrate Performance Testing OpenStack API using Jmeter open source tool. 2. Present key Performance metrics and benchmarks across OpenStack API. 3. Share source repository and ways to contribute for interested test developers. (Session lead is Rohit Karajgi)

Tuesday April 17, 2012 12:00pm - 12:25pm
Seacliff C

12:00pm

Nova-Quantum Integration -> (Quantum Nova API)
Plan: When using Quantum with nova, network hosts/nodes are unnecessary. I propose making the network api configurable via flag (which already exists), and writing a quantum api (nova/network/quantum/api.py) which is able to take the place of both the existing network API and the quantum manager. In this case compute would be making calls directly to Quantum/Melange through the network API. Reviewer Note: the API referred to here is the network.API() class within Nova. It is not referring to the tenant-facing Quantum API. (Session lead is Trey Morris)

Tuesday April 17, 2012 12:00pm - 12:25pm
Seacliff D

12:30pm

Lunch
Tuesday April 17, 2012 12:30pm - 1:30pm
2nd Floor

1:30pm

Lightning talks
Tuesday April 17, 2012 1:30pm - 1:55pm
Ballroom

2:00pm

Communication - IRC, mailing lists, blogs, forums
Review of our community communication media. Does the current situation with IRC and mailing lists (2+ IRC channels, one mailing-list) work well? Should we split into more IRC channels and/or mailing lists? Enforce prefixes on the ML? Should we *split the single ML into developers topics vs. users topics? What should we do with the forums? Should we move to a Q&A site or stay with Launchpad Answers? Those are some of the possible topics, though we can use our session time for any related topic that the session participants are most interested in. (Session lead is Thierry Carrez)

Tuesday April 17, 2012 2:00pm - 2:25pm
Bayview B

2:00pm

Make Nova tenant related data pluggable.
We would like to manage a number of items which are defined per tenant in an alternative central system. The first step is to make the following parts of the system pluggable drivers: - ssh keys - quotas - security group definitions. The default driver would keep this behaviour in the Nova database as per the current implementation, but this would allow alternative drivers to be developed. It would be good to get feedback on this idea and identify any other similar areas. (Session lead is Phil Day)

Tuesday April 17, 2012 2:00pm - 2:25pm
Seacliff AB

2:00pm

Leverage backend store capabilities
A brainstorming session to consider ways for Glance to better leverage the individual capabilities of various backend stores, for example by supporting: - restartable download of partial images - parallel / S3 multi-part image upload - user-driven selection of variable storage QoS, e.g. S3 Reduced Redundancy Storage - propogation of end-user credentials to Swift in order to allow fine-grained storage quotas to be applied - block-level deduplication in Swift to reduce storage footprint of common image data - opportunistic store-specific optimizations, e.g. internal copy between S3 or Swift images (using x-amz-copy-source for S3->S3 or x-copy-from for Swift->Swift) (Session lead is Eoghan Glynn)

Tuesday April 17, 2012 2:00pm - 2:55pm
Seacliff C

2:00pm

IPAM/L3-fwding/NAT/Floating IPs I (was Melange)
I'm putting Melange into Quantum. Quantum is inheriting the Melange database, and we're going to have to expose some of the functionality (ex subnet CRUD) in the Quantum API. Concerns: Auth (Troy Toman has a talk about AuthN/Z -Jason Kolker -Trey Morris (Session lead is Trey Morris)

Tuesday April 17, 2012 2:00pm - 2:55pm
Seacliff D

2:00pm

Growing swift from first install to happy cluster
A full swift install (like is done at Rackspace CloudFiles) is a very sophisticated thing, and works wonderfully; the experience is less good for someone (e.g. an Ubuntu user) that wants to apt-get it on on one machine. For example, currently you need multiple machines (at least as many as you have replicas) - or at least separate instances pretending to be multiple machines - which makes apt-get installation tricky. If you want a single-machine install you have to set replicas=1, but then you can't change this in future without copying all your data in and out manually. The "as unique as possible" blueprint may address this, as might the idea of changing the replica count dynamically. The goal of this discussion is to make it easy to get started with some form of swift install - ideally an apt-get - and then be able to grow from there to a multiple-machine cluster offering real redundancy etc as easily as possible. We will talk about the replica issue, as well as any other issues that people have encountered in practice. (Session lead is justinsb)

Tuesday April 17, 2012 2:00pm - 2:55pm
Marina

2:00pm

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Tuesday April 17, 2012 2:00pm - 3:55pm
Golden Gate

2:30pm

Support for ARM processors.
During the Diablo cycle work was done to make nova-compute work for ARM processors. We need to finish the work that we started to make ARM a first class supported architecture for folsom. (Session lead is Chuck Short)

Tuesday April 17, 2012 2:30pm - 2:55pm
Seacliff AB

2:30pm

DEVSTACKpy and what it can do for u!
Come learn DEVSTACKpy! The completely Python developer OpenStack deployment and development environment. Run through description of: 1. Its features 2. Its enhancements over devstack.sh 3. How it can be easily used and configured 4. Documentation and code examples Brainstorm the following: 1. Future improvements a. Multi-node dev. installs? b. Rollbacks and more transactional like installs? c. How to keep the different package versions in sync? 2. Any other feedback/questions Demo the following: 1. Simultaneous OpenStack install on Ubuntu 11.10, RHEL 6, Fedora 16 a. Followed by a starting of the base OpenStack components and then an uninstall (on all 3 distros) We can go through the following as well: https://launchpad.net/~devstackpy (our small dev. group) https://github.com/yahoo/Openstack-DevstackPy (if anyone wants to see the code) http://readthedocs.org/docs/devstackpy/ (docs for those interested) (Session lead is Joshua Harlow)

Tuesday April 17, 2012 2:30pm - 3:25pm
Bayview B

3:00pm

Config drive improvements
As discussed https://lists.launchpad.net/openstack/msg07581.html , config-drive is less OS agnostic than we would like it to be. I think that a discussion would lead to a more general purpose mechanism for communicating data from platform to guest without a network. (Session lead is Scott Moser)

Tuesday April 17, 2012 3:00pm - 3:25pm
Seacliff AB

3:00pm

python-swiftclient
Currently the CLI and the library swift.common.client is directly included with Swift which make it hard to integrate with application like Glance for example. I'd like to discuss to extract the CLI and the library to its own package like for other client libraries. (Session lead is Chmouel Boudjnah)

Tuesday April 17, 2012 3:00pm - 3:25pm
Marina

3:00pm

Glance Image Replication
Let's talk about replicating images between deployments in Glance! (Session lead is Brian Waldon)

Tuesday April 17, 2012 3:00pm - 3:55pm
Seacliff C

3:00pm

IPAM/L3-fwding/NAT/Floating IPs II (was L3 API)
A presentation was made on this topic during the Essex Summit, and subsequently a Blueprint and API specs were published to the community. We would like to carry forward this discussion. More information on this wiki: http://wiki.openstack.org/quantum-l3 (Session lead is Sumit Naiksatam)

Tuesday April 17, 2012 3:00pm - 3:55pm
Seacliff D

3:30pm

Bug handling improvements
We use Launchpad to handle OpenStack bugs. This session will discuss further improvements and changes in the way we use various status / importance / tags, see what we can do to encourage more people to participate into triaging filed bugs. We'll also discuss the fate of abandoned "Wishlist" items that clog some of the bug lists, and how much those should just be turned into blueprints or items on a wiki page. (Session lead is Thierry Carrez)

Tuesday April 17, 2012 3:30pm - 3:55pm
Bayview B

3:30pm

Guest networking
There needs to be a standard around how network information get injected into images. Some possibilities are the following: 1. Usage of https://fedorahosted.org/netcf/ (or its client library) inside of nova when injecting into the image 2. Usage of https://fedorahosted.org/netcf/ inside a VM (nova would write the agnostic format into the VM) and having software inside the VM on firstboot interpret that format (cloud-init??) 3. Using the current ubuntu network format that is already written, but change where it is written to an agreed upon location and perform #2 (the ubuntu format would then be the "agnostic" format) 4. Do nothing and cry. (Session lead is Joshua Harlow)

Tuesday April 17, 2012 3:30pm - 3:55pm
Seacliff AB

3:30pm

Swift Cluster Monitoring With StatsD
This presentation will outline the rationale and implementation of Swift cluster monitoring using StatsD. We will also discuss how this monitoring can integrate with other infrastructure and tools (StatsD server, Graphite, etc.). (Session lead is Darrell Bishop)

Tuesday April 17, 2012 3:30pm - 3:55pm
Marina

4:00pm

Coffee Break
Tuesday April 17, 2012 4:00pm - 4:30pm
2nd Floor

4:30pm

Nova use of read-only/public glance server
Currently, if nova is configured to use a glance instance, it basically assumes it can also write to that glance (at least for snapshots of the instance). This means that public glance servers aren't really feasible. I'd like to be able to run a public glance server, and configure nova installations to offer images presented there as runnable. Some of the things that would need to be covered here: * local caching of remote images * local storage of snapshots of remote image (Session lead is Scott Moser)

Tuesday April 17, 2012 4:30pm - 4:55pm
Seacliff C

4:30pm

VM State Management (take 2)
At the Essex design summit I presented an idea to limit the operations on a Instance based on the combination of vm_state and task_state. We tried to make the model quite restrictive to eliminate as many race conditions as possible (for example rebuilds occurring in the middle of a rebuild). However having run with this model for some time we have found it necessary to allow more operations that we at first envisaged to cope with Instances that get “stuck” in a particular state (for example is a service is restarted whilst a rebuild is in progress the Instances can get stuck in “Rebuilding-Spawing”). Only allowing delete in this case (which is the operation which always leads to a deterministic end state no matter what it’s paired with) isn’t a great user experience, but allowing other operations (such as a further rebuild) creates more potential race conditions that have to be handled in the code. I therefore propose moving to a model which can block operations on a combination of vm_state, task_state, and updated_at – leading to three possible rules: i) Operations which are always allowed ii) Operations which are always blocked iii) Operations which are allowed only if now()- instance[updated_at] is greater than some thresholds (which can be different for each operation) So for example, for an Instance which has the state Active-None Start, Unpause, Resume, UnRescue would always be blocked Other operations would always be allowed Whereas for an instance which has the state Active-Rebooting: Start, Unpause, Resume, and Unrescue would always be blocked Terminate, Reboot, and Rebuild would only be allowed if the now()-instance[updated_at] is greater than 3 minutes I believe this consideration of “how long since the last time the state changed” comes very close to the symanatic of limiting each instance to only one concurrent operation whilst still being safe against failed operations and without having to make major changes to the overall concurrency model. I would also like to explore with the community the degree to which these rules should be configurable vs hard coded (for example some might see a need to be able to configure to always allow operations such as delete which have a billing impact). (Session lead is Phil Day)

Tuesday April 17, 2012 4:30pm - 4:55pm
Seacliff AB

4:30pm

API extensions: the good, the bad, and the ugly
API extensions are a standard way to extend the functionality of APIs and expose them to users. However, some of the OpenStack API extension mechanisms are perceived as clunky and error-prone. I'd like to use this session to brainstorm ways of creating and supporting an extension mechanism that would be simpler and easier to use; if our proposal is good enough, we can implement it as part of Glance's v2 API (per the Glance PTL). (Session lead is Glen Campbell)

Tuesday April 17, 2012 4:30pm - 5:25pm
Bayview B

4:30pm

Security Groups & Firewalling
Discuss proposal for security groups exposed via the Quantum API. Security group membership would be on a per-port basis. We will also brainstorm around additional firewalling capabilities, including a more generic ACLing capability. (Session lead is David Lapsley)

Tuesday April 17, 2012 4:30pm - 5:25pm
Seacliff D

4:30pm

Implementing a Cloud Storage Standard for Swift
There is growing momentum behind a cloud storage standard that can level the playing field and expand the market for cloud storage overall by removing the friction of moving from cloud to cloud. As Swift gains increasing functionality from the broad community of support, it just makes sense to use a standard API to guide the interface to those features. The Cloud Data Management Interface (CDMI) is a cloud storage standard for both the datapath to cloud stroage as well as a standanrd for managing the data once it is in the coud. CDMI is being implemented by commercial and academic organizations and there is a blueprint for incorporating an implementation in OpenStack. This talk will look at the existing features of Essex and planned features for Folsom, comparing them to the features that are already standardized in CDMI. The talk is intended for a technical audience that is already familiar with the workings of cloud storage in general and Swift in particular. (Session lead is Mark Carlson)

Tuesday April 17, 2012 4:30pm - 5:55pm
Marina

5:00pm

Resource rationing in Glance
Brainstorm on whether Glance needs to protect itself from inadvertent DDoS'ing by throttling eager clients, and also whether fairness should be enforced by rationing resource usage. The obvious approach would be to model this on nova rate-limiting, possibly leveraging the distributed mechanism provided by Turnstile in the case where the Glance API service is horizontally scaled. However Glance may have additional requirements, for example weighting GET calls on the size of the image returned (which can vary by several orders of magnitude). Also the nova quota/quota-classes API extensions could be useful to re-use, though again Glance may have some special requirements, for example weighting resource usage depending on the backend store in use (limited local filesystem space may for example attract a higher weighting than effectively infinite S3 storage). (Session lead is Eoghan Glynn)

Tuesday April 17, 2012 5:00pm - 5:25pm
Seacliff C

5:00pm

Nova-Orchestration
Need for an orchestration layer/ service has been visited and talked about during Essex summit (http://etherpad.openstack.org/orchestration). However, Orchestration couldn't be materialized during Essex. This session is to revisit the need, identify the gaps, brainstorm and review design options and come up with a plan to implement the same. Summary from previous Summit: Nova needs to be able to monitor and manage long-running transactions. For example, when creating 1,000 servers, it may be necessary to create a build plan, retry failed builds, refresh the plan, and possibly even roll back the entire transaction if it cannot be completed rapidly enough. To do this, an orchestration service is proposed that can monitor workflows around Nova events and transactions. (Session lead is Sriram Subramanian)

Tuesday April 17, 2012 5:00pm - 5:55pm
Seacliff AB

5:30pm

Demo of Testing OpenStack with Tempest/Devstack
I will demonstrate how to test OpenStack easily with devstack and the Tempest integration suite. Intended to explain the relation between Tempest and devstack with regards to easy testing of OpenStack environments, walk through how Tempest is structured and provide information for contributors interested in participating in the QA team through Tempest. Tempest project: * https://launchpad.net/tempest * https://github.com/openstack/tempest Devstack: * http://devstack.org * https://github.com/openstack-dev/devstack Written walkthrough of what will be demonstrated: * http://www.joinfu.com/2012/03/testing-essex-rc1-with-devstack-and-tempest/ (Session lead is Jay Pipes)

Tuesday April 17, 2012 5:30pm - 5:55pm
Bayview B

5:30pm

Glance Backend Image Autodetection
A brainstorm session to consider ways for glance to automatically detect and register images on a backend. This goes somewhat hand-in-hand with NFS backed images, because there are use cases for organizations that already handle image management for bare-metal or other virtualization and should not have to manually register each image when they have hundreds. (Session lead is ChristopherMacGown)

Tuesday April 17, 2012 5:30pm - 5:55pm
Seacliff C

5:30pm

Scaling the Quantum API
As production usage of Quantum ramps up, we need to make sure the API will scale. We would like to gain agreement on support some of the following concepts in the API - Add distributed rate limiting - Add API caching - Possible support for bulk operations - Possible support for some orchestration functions to eliminate the overhead of multiple calls - Opportunities to make the API more asynchronous (Session lead is Troy Toman)

Tuesday April 17, 2012 5:30pm - 5:55pm
Seacliff D

5:30pm

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Tuesday April 17, 2012 5:30pm - 5:55pm
Golden Gate

6:30pm

HP Party at the Ferry Building

Expect surprise guest entertainment and great food!

6:30pm – 8:30pm:            Networking Reception and Heavy Appetizers
8:30pm – 9:30pm:            Exclusive Concert
9:30pm – 10:00pm:          Afterglow and Conclusion


Tuesday April 17, 2012 6:30pm - 10:00pm
Grand Hall 1 Ferry Building, San Francisco, California
 
Wednesday, April 18
 

9:00am

Common image properties
When launching an instance from code, the code needs to be able to pick a suitable image. Today, there is no easy way to do that, particularly when targeting multiple OpenStack clouds. Some clouds expect us to parse the name, some use metadata (different metadata for each cloud). I am proposing that the public clouds agree a set of common image properties. What they are doesn't really matter, but here's something I would consider sane: openstack.org:os_distro=debian.org openstack.org:os_major=6.0 openstack.org:os_minor=4 openstack.org:updated_through=20120401 I am not proposing that we do this through code. It should be possible for a public cloud to implement this without any code changes. (Session lead is justinsb)

Wednesday April 18, 2012 9:00am - 9:25am
Seacliff AB

9:00am

Puppet/OpenStack
This session will serve as a high level introduction to Puppet, focused on how it complements OpenStack. The session will cover the following: - A high level introduction of Puppet - An overview of our project to create Puppet modules for building and managing OpenStack - A call for the consolidation of existing OpenStack/Puppet modules into a master set (Session lead is Dan Bode)

Wednesday April 18, 2012 9:00am - 9:55am
Bayview B

9:00am

Horizon Folsom Roadmap
A deep-dive working group session where key blueprints for Horizon will be discussed, and implementation/API details can be worked through. Expected topics include "workflows", consistent ajax data handling, and tenant deletion among others. (Session lead is Gabriel Hurley)

Wednesday April 18, 2012 9:00am - 9:55am
Seacliff C

9:00am

Security improvements in Nova for Folsom
Folsom should be high-security :) What security architectural improvements could we push during the Folsom timeframe ? This includes discussion of encrypted internal communication and further improvements to make the rootwrap less permissive. (Session lead is Thierry Carrez)

Wednesday April 18, 2012 9:00am - 9:55am
Marina

9:00am

Nova parity for open-source quantum plugin(s)
In folsom, Quantum will become core, and our target is that it will become the default model for Networking with OpenStack (existing network managers will likely continue to exist for backward compat). For this to succeed, we'll need at least one fully open source plugin that can be deployed in a manner similar to any of the three modes of nova-network. This will require non-trivial work, as most network related capabilities are being deprecated in Nova. This includes support for: - IPAM integrated with Quantum (port of Melange) - DHCP injection based on IPAM config (work done here by locaweb + nicira) - security groups (port iptables implementation from nova layer) - L3 + NAT implementation (port iptables logic from nova-network) - "provider networks" (i.e., mapping to a specific vlan in the physical network) - provider firewall (port firewall driver from nova virt layer) - multi-host support for L3 + DHCP - VPN (need input from Vish as to whether this is needed). - Floating IPs In many cases this can be helped by lifting existing code from nova-network or the firewall driver in nova's virt layer. This code should be structured as We probably also want to include plans for various improvements to the existing plugins: - removing polling and instead using something like rabbitmq - restructure code to allow better sharing of code across agents. We also need to consider how we might offer a mechanism to migrate nova networking configuration and state to Quantum. Note: we'll also have to think about different virt layers: libvirt is pretty straightforward, but xenserver uses some interesting tricks witha guest VM to do iptables filtering + L3. ESX is obviously another interesting question. (Session lead is dan wendlandt)

Wednesday April 18, 2012 9:00am - 10:25am
Seacliff D

9:00am

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Wednesday April 18, 2012 9:00am - 10:25am
Golden Gate

9:30am

Efficient metering for Nova and Swift
The goal of this session is to review a proposed architecture blueprint to implement a proper metering mechanism within Nova and Swift and make sure we gathered the appropriate requirements. (Session lead is Nick Barcet)

Wednesday April 18, 2012 9:30am - 10:25am
Seacliff AB

10:00am

Colony
This session will be a presentation to review what has been done in colony so far and the directions the community could take going forward. Colony is a project which federate cloud services. The fisrt target is the federation of object services, like swift. A very brief introduction to colony is at https://github.com/nii-cloud/colony. Colony can be a project which makes Openstack a strong player in enabling federated cloud ecosystems. (Session lead is Shigetoshi Yokoyama)

Wednesday April 18, 2012 10:00am - 10:25am
Bayview B

10:00am

Horizon and eco-system project integration
Discuss the best way to integrate eco-system projects such as Red Dwarf, Atlas, and others in a generic and scalable way. (Session lead is Devin Carlen)

Wednesday April 18, 2012 10:00am - 10:25am
Seacliff C

10:00am

Ops pain points
What are the pain points for operations staff maintaining openstack installations? There have been some small improvements in essex in this regard (for example trying to log the instance UUID for errors). What else can we do to make it easier for ops people to diagnose faults and explain the state of instances which have gone wrong? (Session lead is Michael Still)

Wednesday April 18, 2012 10:00am - 10:25am
Marina

10:30am

Coffee Break
Wednesday April 18, 2012 10:30am - 11:00am
2nd Floor

11:00am

Java bindings for OpenStack
Presentation of the Java bindings for OpenStack that Luis & I have developed, along with some initial applications (a CLI, GWT, Jenkins plugins, Java 7 provider etc) Also a discussion of some of the problems we encountered, trying to use the XML schema etc from a strongly typed language, and possible implications (should we continue to support XML?) (Session lead is justinsb)

Wednesday April 18, 2012 11:00am - 11:25am
Bayview B

11:00am

Autobackup When Deprovisioning
On virtual machine deletion (termination) there needs to be a way to safely transmit the state of that virtual machine (disk state to begin with, memory state maybe later in the future) to a location for a given period of time. This is useful for debugging, tracking and auditing (and even backing up) of terminated machines for a given period of time. Most of this functionality already exists in openstack but there are some slight additions that need to be added. This type of feature should be available for the people that control the cloud as well as users who run on the cloud (as a private company that runs ontop of the cloud may want the same functionality for there own uses). (Session lead is Joshua Harlow)

Wednesday April 18, 2012 11:00am - 11:25am
Marina

11:00am

Quantum System Test
This session will describe the integration testing infra available to ensure that we have a solid process to ensure that unit tested Quantum commits pass basic integration tests. (Session lead is Debo~ Dutta)

Wednesday April 18, 2012 11:00am - 11:25am
Seacliff D

11:00am

High Availability in OpenStack
A brainstorm session for discussing ideas to achieve true high availability for OpenStack components. For a few infrastructure services, this is straightforward as solutions already exists: - MySQL - RabbitMQ - also load balancing (not really HA, but related) Other services are entirely stateless, and upstart integration exists, so we can just tuck them under a cluster manager and get HA pretty much for free. Then there are some where integration with cluster management does not exist yet, but should be relatively easy to do, specifically in Glance. Then there are some tricky ones, such as nova-volume. One possible option that may come in handy is deeper Ceph integration (with RBD). And finally, there's HA for guests, for which there's a blueprint. (Session lead is Florian Haas)

Wednesday April 18, 2012 11:00am - 11:55am
Seacliff AB

11:00am

Nova Plugin Framework design workshop

uring the 'Split Nova and make it more pluggable' session there was strong support for addition of a plugin framework in Folsom. This session will involve pulling apart my proposed design and building up a better one. http://etherpad.openstack.org/NovaPlugin

Join Andrew Bogott.


Wednesday April 18, 2012 11:00am - 11:55am
Golden Gate

11:00am

Building on Horizon
A deep-dive session on how to build your own Dashboards, Panels, DataTables, TabGroups, and more, plus you'll learn how to integrate them all into your Horizon installation. This is the nuts and bolts for anyone who wants their project to work with the OpenStack Dashboard! (Session lead is Gabriel Hurley)

Wednesday April 18, 2012 11:00am - 12:25pm
Seacliff C

11:30am

Support for SPICE remoting
SPICE is an open source alternative to VNC. It promises much better performance and capabilities than VNC, and could enable desktops-as-a-service with OpenStack. I have a quick POC of SPICE integration here, which attracted a lot of interest: https://review.openstack.org/#change,5319 I hope to evaluate SPICE further (to see if it lives up to the hype), discuss my findings, and work with the community to add support for SPICE to OpenStack. A potential plan of action is: 1) Push (insecure) code that allows administrators to turn on SPICE and evaluate it. 2) Develop a gateway like the novnc gateway that allows secure client access if #1 has good results. 3) Native client port of SPICE C code? (Session lead is justinsb)

Wednesday April 18, 2012 11:30am - 11:55am
Marina

11:30am

Quantum Automated Testing
There's always room for improvement in a project's approach to automated testing, especially when that project interacts with a system-level component like networking. This session will offer space for discussion of people's experience testing Quantum, and hopefully generate consensus on what works, what doesn't, and what could be done to improve both the quality and maintenance cost of the test suite. (Session lead is Maru Newby)

Wednesday April 18, 2012 11:30am - 11:55am
Seacliff D

11:30am

How to track contributions to the community
Over 300 people from over 50 companies committed code to OpenStack during its, more than 150 companies are involved in the project, thousands of bugs tracked, tens of thousands of email messages, wiki edits, forum discussions... We'll present the current structure of the data collected from community contributions, the standard reports and a first version of a public portal to analyze such data. We should leave enough the time to discuss how to increase size and scope of the community metrics system of CpenStack. What part of this data is useful information for the OpenStack community? Are we collecting the right data? From the business perspective, what kind of data would be useful to track? (Session lead is Stefano Maffulli)

Wednesday April 18, 2012 11:30am - 12:25pm
Bayview B

12:00pm

Base Packaging Guidelines
Devstack is a great template for a development environment, but we don't have example packaging for production usage. This means that each downstream package has to roll-their-own, and I've also observed some pretty big gotchas, like mixing user configuration into the routing files (Bug #956709) If we had an example production configuration: 1) We would identify and fix any packaging issues (early) 2) Packagers wouldn't each have to solve these issues (duplication of effort, fragmentation issues) 3) We would hopefully converge on the same basic setup - e.g. service names, log file locations etc, which would make documentation and problem diagnosis much easier 4) It should be practical to run a production deployment from source I hope that the various OS packaging efforts can contribute their problems & solutions, that we can pick the best, and produce either a document or automated deployment code. Hopefully this can then help downstream packagers. I've been doing some investigations here: https://github.com/justinsb/openstack-simple-config (Session lead is justinsb)

Wednesday April 18, 2012 12:00pm - 12:25pm
Seacliff AB

12:00pm

Host matchmaking
The ZeroMQ RPC driver introduces a concept of pluggable matchmaking for bare topics providing a "get_workers" method. This can be used to find a peer to communicate with, a la "compute.host" rather than sending to a bare topic such as "compute". A similar requirement exists in the scheduler when it pulls a list of all hosts from the database (which feeds from RPC). These efforts should be condensed. Having pluggable matchmaking can immediately reduce the scheduler's dependence on the database and has clear advantages for distributed RPC (and potentially for (de)centralized RPC as well). We should brainstorm requirements, where this fits in, etc. (Session lead is Eric Windisch)

Wednesday April 18, 2012 12:00pm - 12:25pm
Marina

12:00pm

Quantum in the Data Center
This session explores enhancements to Quantum enabling it to better integrate with existing data center network environments. We will outline some of the challenges identified while exploring the use of Quantum with oVirt for enterprise virtualization, and brainstorm on potential solutions. Issues identified include incorporating existing (virtual) networks, handling a mix of networking technologies, and modelling asymmetric connectivity. (Session lead is Robert Kukura)

Wednesday April 18, 2012 12:00pm - 12:25pm
Seacliff D

12:00pm

How to Get OpenStack - Wiki updates

Installation information on the OpenStack wiki is currently unorganized, and sometimes very much out of date. Also, it is not at all clear how to get OpenStack. Plan is to create a wiki page summarizing the varios installers, distros and packages, ppa's, and commerical distribution to enable one stop shopping.

Join Mike Pittaro.


Wednesday April 18, 2012 12:00pm - 12:25pm
Golden Gate

12:30pm

Lunch
Wednesday April 18, 2012 12:30pm - 1:30pm
2nd Floor

1:30pm

Lightning talks
Wednesday April 18, 2012 1:30pm - 1:55pm
Ballroom

2:00pm

Repose Status

Give a brief status of the Repose project. Discuss new features.

Join Jorge Williams.


Wednesday April 18, 2012 2:00pm - 2:30pm
Golden Gate

2:00pm

FederatedZones, meta-affinity with ServiceCatalogs
At their most abstract, zones are a meta-concept layered over independent installations of OpenStack services. Whether nova cells, or swift regions, whatever quantum calls them, or federated instances of glance, a base zone need not be aware of the infrastructure, data, or implementation of their peers. After authenticating with Keystone, you receive back a service catalog containing the end-points for each service. Since we return endpoints in a list, we already expose a way of handling zones in a simple, effective manner. By making Keystone's ServiceCatalog better, we can simplify distributed scheduling in Nova by doing it at a meta layer, simplify swift cross-region replication, expose glance's federated image repositories within nova more easily, and also potentially give everyone a pony. (Session lead is ChristopherMacGown)

Wednesday April 18, 2012 2:00pm - 2:55pm
Seacliff AB

2:00pm

Reddwarf PaaS building blocks
As more companies move their applications and data to the cloud, it is becoming increasingly difficult to manage and maintain PaaS systems on default virtual servers. RedDwarf aims to simply PaaS management in the cloud while providing a model for extensible service deployment that will be used to deliver not only database services, but also other services in the future. In this session you will get a chance to hear about our progress and future plans and learn how you can become active in the community. Reddwarf has been significantly refactored to bolt on top of nova properly and treat nova as a blackbox system. It has its own state information, its own database, and its own communication infrastructure. I would like to brainstorm on its current state, and how to best create the building blocks of a generic PaaS solution that can be used to build out multiple services. We can talk about the state of the current Reddwarf redux, and where to go with it in the future. Code/Docs can be found here. https://github.com/hub-cap/reddwarf_lite Integration (redstack) is here. https://github.com/hub-cap/reddwarf_lite-integration Client binding here. https://github.com/hub-cap/python-reddwarfclient (Session lead is Michael Basnight)

Wednesday April 18, 2012 2:00pm - 2:55pm
Bayview B

2:00pm

OpenStack UX
This session will begin with an introduction to the currently proposed design process, including a quick overview on the Human Interface Guidelines document. The rest of the time will be used as a working session to dive in and discuss user experience design for the Horizon Dashboard. (Session lead is Paul Tashima)

Wednesday April 18, 2012 2:00pm - 2:55pm
Seacliff C

2:00pm

Nova EC2 compatibility
Nova has some EC2 compatibility, but could be better. This session will look at how to support users coming from Amazon to OS: what is most needed, and where to add it. This will also include discussion on whether the ec2-api should be managed inside of nova or via an external component. (Session lead is Martin Pool)

Wednesday April 18, 2012 2:00pm - 2:55pm
Marina

2:00pm

L2 & L3 Network Services Insertion
Library for VM-based service insertion (Edgar) Provide a library to insert "In-Path" and "Out-of-band" layer 2 & Layer 3 network services such as firewalls, LB, Routers or WAAS. There is no need for new APIs at this moment because the insertion of the services will be calling current APIs. This is an extra functionality for SPs who will be able to instantiate services as VMs in the network with or without L3 configuration. Services-Insertion using non-core Quantum API Extensions (danwent) Basic idea is how someone can use Quantum API extensions to provide a service that builds on top of the existing capabilities of the plugin. The idea is that you run the core plugin, but you can also run one or more services, which are essentially python modules that are loaded Quantum startup and that register API extensions that do not conflict with the core Quantum APIs (e.g., a module might introduce a new /load-balancers URL space). The logical network entities exposed by the services can use admin interfaces to associate interfaces (e.g., load balancer interfaces) into Quantum network ports. This generic service can be written in a way that it is compatible with multiple plugins, and can be enabled independently of any plugin. (Session lead is Edgar Magana Perdomo)

Wednesday April 18, 2012 2:00pm - 2:55pm
Seacliff D

3:00pm

Making the configuration of openstack easier
As OpenStack gains broader adoption, it will have to become easier to manage and deploy. I would like to lead a discussion about things that could be done to make OpenStack easier to configure (like .d directories) Hopefully we could use this session to file some tickets or create a blueprint. (Session lead is Dan Bode)

Wednesday April 18, 2012 3:00pm - 3:55pm
Seacliff AB

3:00pm

Cloud Foundry, BOSH, and CPIs like OpenStack
Presenter: Vadim Spivak and friends. Last minute addition to the Design Summit. Vadim will provide an technical overview of Cloud Foundry BOSH, the newest open source project from the Cloud Foundry team at VMware. The talk will be illustrated with ample code samples and will include a deep dive into the Cloud Provider Interface (CPI), the API in BOSH used to abstract the underlying infrastructure, including, for example, OpenStack. (Session lead is Lloyd Dewolf)

Wednesday April 18, 2012 3:00pm - 3:55pm
Bayview B

3:00pm

Horizon and Quantum Integration
Develop a plan for making Quantum a first class citizen within Horizon for Folsom with the goal of providing a great user experience for creating and managing virtual networks. (Session lead is Devin Carlen)

Wednesday April 18, 2012 3:00pm - 3:55pm
Seacliff C

3:00pm

Bridging the HPC Gap using OpenStack
The focus of this session is to discuss the current state of High Performance Computing using OpenStack and idenify development priorities and resources to further HPC-class workloads in the next release. Topics include cluster support and topology aware scheduling, I/O issues such as Infiniband support, RDMA-over-*, latency-sensitive I/O, security, etc., large memory guest VMs and compute accelerator access (GPUs, FPGAs, etc.), large data management challenges such as parallel file systems, and integration with HPC ecosystems and tools. (Session lead is BrianSchott)

Wednesday April 18, 2012 3:00pm - 3:55pm
Marina

3:00pm

Placeholder: Horizon Track discusses Quantum
There will be no Quantum session in this slot, as most Quantum team members will be attending the Horizon track session on Horizon exposing Quantum functionality. (Session lead is dan wendlandt)

Wednesday April 18, 2012 3:00pm - 3:55pm
Seacliff D

4:00pm

Coffee Break
Wednesday April 18, 2012 4:00pm - 4:30pm
2nd Floor

4:30pm

Instrumenting OpenStack
Gathering of metrics (availability, performance) and reporting in a standard fashion should be part of OpenStack. Best practice sensors should sample the metrics and provide alarms for issues which could cause service impacts. Posting of these alarms to a monitoring system should be based on plug ins. cluster health and monitoring: * each swift service has a "healthcheck" call that can be used by a monitoring system * it would be a big improvement for deployers if each openstack service provided healthcheck apis, and * these were based on expert knowledge of what is supposed to be happening inside This would also insulate deployers from changes in the code that might impact what it means to be running properly. Closely related: providing standardized logging and documenting error conditions. Then, various tools could be applied to the logs (splunk, syslog, logstash, etc.). (Session lead is Duncan McGreggor)

Wednesday April 18, 2012 4:30pm - 4:55pm
Seacliff AB

4:30pm

CLI / Client Rewrite
The CLI and python client needs some love. Lets rewrite it using the cement2 framework to provide a rich CLI experience and create a true python client for integration code that the cli will then use to provide the connectivity. (Session lead is Jason Koelker)

Wednesday April 18, 2012 4:30pm - 4:55pm
Seacliff D

4:30pm

CI and Developer Infrastructure Roadmap for Folsom
We made some huge strides forward for Essex, and would like to continue the trend for Folsom. We'll start by presenting an overview of what we accomplished from a tooling and automation perspective in the last cycle, and then walk through what our current plans are for the next cycle. More importantly though, we'll do that reasonably quickly so that we can solicit feedback on what people need. (Session lead is Monty Taylor)

Wednesday April 18, 2012 4:30pm - 5:25pm
Bayview B

4:30pm

Horizon and Swift Recon Integration
Develop a plan for exposing Swift Recon statistics from within the Horizon system administrator panel. We have an opportunity to gather and present a variety of useful metrics and operational health of a Swift deployment. (Session lead is Devin Carlen)

Wednesday April 18, 2012 4:30pm - 5:25pm
Seacliff C

4:30pm

Per-instance Package Configuration via Nova
The attached design proposes an API which would allow instances to be customized via puppet, chef, etc. using a Nova API extension. Several people have expressed interest in this feature; several others have argued that it should be managed outside of the Nova API. Topics of discussion: - Does anyone else want this? - Is adding a Nova API interface for this feature useful? Should it be part of nova-core or an extension? - Is the proposed blueprint missing necessary features or flexibility? - Who will write this and/or who will write tool-specific drivers? - In order to support this, what features do we require from the hypothetical/future guest-agent tools? (Session lead is Andrew Bogott)

Wednesday April 18, 2012 4:30pm - 5:25pm
Marina

4:30pm

Unconference sessions

Proposals made and decided upon live in real-time by the participants.


Wednesday April 18, 2012 4:30pm - 5:55pm
Golden Gate

5:00pm

OpenStack and Operations: Getting Real
Operations intersects every project and subproject in OpenStack. Yet Ops gets very little interest and attention from core developers who are quite understandably focused on adding necessary features to OpenStack. This discussion will brainstorm on the following: * how can we make Ops more attractive so that folks want to be more involved? * Ops is currently a subteam under nova; does it belong here? * should we create some sort of operational project that spans all projects? * perhaps a project manager? and * the PTLs would serve as a kind of board? * this project could gather information from live deployments and operations in the wild, and present it to the PTLs on a regular basis * should we integrate with professional groups like the IEEE? (e.g., for specifying some standards for cloud orchestration, etc.) (Session lead is Duncan McGreggor)

Wednesday April 18, 2012 5:00pm - 5:55pm
Seacliff AB

5:00pm

Quantum Summit Wrap-up & Roadmap
This will serve as a change to summarize + review plans resulting from the summit, identify any crucial gaps, talk about relative priorities, dependencies, and development timelines. (Session lead is dan wendlandt)

Wednesday April 18, 2012 5:00pm - 5:55pm
Seacliff D

5:30pm

I18N in OpenStack
As of Essex only Nova and Glance use message translations, and the translation effort decreased significantly since Cactus. Do we really want translations, for messages that are usually sysadmin-facing ? If yes, can we standardize I18N across all core projects ? (Session lead is Thierry Carrez)

Wednesday April 18, 2012 5:30pm - 5:55pm
Bayview B

5:30pm

Guest agents support and implementation
One being involved into multiple commercial OpenStack deployments can’t help noticing that one of the most frequent requests from customers is the support of Windows as a guest operating system. At the same time customers expect this support to be somewhat smart compared to the *nix like systems support. Currently there seems to be no better non-hackish solution than guest agents once proposed and implemented by the Rackspace company for Xen-based OpenStack installations. The guest agent is a background service which runs within the guest operating system, receives and executes commands issued by a cloud user via the CLI or web-interface. The sphere of guest agent applications is not limited with Windows-related features support, it includes lots of guest side magic for other guest types as well. Scope of possible applications for guest agents - Instance access recovery = Resetting administrator’s password in a secure manner = Resetting networking configuration to default one - Guest-side preparation of an instance for further snapshotting (in particular actual for Windows instances) - Guest side software installation and configuration (serving as a Chef/Puppet bootstrap) - Block device management = Optional creation of file system on newly attached block devices = Resizing of system partition on instance spawn according to the flavor used - Advanced monitoring of an instance state = Collection of some sophisticated parameters (e.g. disk or memory usage) which can’t be obtained from outside - Updating of guest side PV drivers for a specific hypervisor (e.g. Xen Guest tools) - File system-agnostic file injection Proposed agenda for brainstorming - Responsibilities and functions of guest-side agent - Security considerations, restricting access to instances - Implementation details = Programming language, toolset and packaging for various target OS - OpenStack and guest agent interoperation details = Communication channels in case of different hypervisors = Choosing an appropriate communication model (pull/push, persistent connection/polling etc.) = Definition of low-level communication protocol details - OpenStack-side support for guest agent functions = Existing management command support (like nova set-root-password) = Additional management commands definition and implementation = Extension of Nova API (if required) = Dashboard extension (Session lead is Artem Andreev)

Wednesday April 18, 2012 5:30pm - 5:55pm
Marina

6:00pm

Cloud Foundry BOSH + Openstack Hackathon

Ever wanted to get involved with some coding to make PaaS a reality on OpenStack? Join us for a special Cloud Foundry code deep-dive and hackathon that follows the Design Summit & Conference regular sessions. This is your chance to master Cloud Foundry and BOSH while contributing to the effort of getting Cloud Foundry running on OpenStack. The goal of the Cloud Foundry hackathon is to create a solid foundation for creating integrations between Cloud Foundry BOSH and OpenStack. Of course you'll come for the code, but food and drinks will be on offer. 


Wednesday April 18, 2012 6:00pm - 9:00pm
Dev Lounge Hyatt Regency

8:00pm

Nebula OpenStack Conference Launch Party

http://openstackconference-launchevent-apr2012.eventbrite.com/

Join Nebula as we transition from the Design Summit into the Conference with an evening of music and entertainment with your fellow colleagues.

Nebula welcomes all OpenStack conference participants to the official conference kick-off party at Public Works situated in San Francisco’s Mission District.

We’ll be pouring drinks, serving food, enjoying some live music, seeing some amazing local performance artists, and checking out some eclectic art between 8pm and midnight on Wednesday April 18th .

Just hop on a trolley shuttle outside the Hyatt Embarcadero between 8 and 12. Please come by and relax with us.


Wednesday April 18, 2012 8:00pm - 11:59pm
Public Works 161 Erie Street (off Mission St. between Duboce & 14th St.)