2014 Events

Speaker: Tom Tromey

In recent years GDB has undergone a renaissance, adding Python scriptingand other cool new features. This talk will show you how to customizeGDB for your application and your debugging needs. We’ll go into depthabout pretty printing, stack trace filtering, and writing new commands;and will also discuss writing GUIs and other tools inside GDB. Finally,we’ll cover other interesting and useful GDB projects.

Speaker: Daniel Berrange

Entangle is an open source project that provides a Linux desktopapplication for “tethered shooting”. It uses the ligphoto2 library totrigger the camera shutter, preview shots via the camera’s “live view”capability, download captured images and access all the live camerasettings. It is useful for a variety of use cases including studio modelshoots, macro still life, stop motion animation shoots,astro-photography and more.

The talk will be targeted at Linux users & developers who arephotographers interested in any of the aforementioned use cases.

Speaker: Federico Simoncelli

This session will cover the current status of integration between oVirtand the OpenStack image repository (Glance), analyzing the motivations,the low level implementation (including Keystone authentication) and theideas for the future. The presentation will include also an ample partdedicated to the future work and ideas to introduce the integration withCinder (the OpenStack volume manager).

• Introduction to oVirt Storage Architecture
• Glance Integration Motivations
• Glance Integration Deep Dive
• Ideas and Future Work on Glance Integration
• Future Integration with Cinder
• Roadmap

Speaker: Daniel Berrange

In the 8+ years since it was founded, the libvirt project grown tobecome the leading open source API for the management of virtualizationhosts, with a strong focus on supporting the open source virtualization& container technologies, KVM, QEMU, Xen and LXC. Many people working inthe open source virtualization management space already have anunderstanding of the core features and architecture of libvirt. Thistalk will thus focus on a selection of recently developed features andof some of the other important, but less well known, features oflibvirt.

The talk will be targeted at virtualization application developers usinglibvirt, with a bias towards those using KVM or LXC. At the end of thetalk the audience will better understand how to take advantage oflibvirt for their development needs.

Speaker: Ohad Levy

Foreman is a complete lifecycle management tool for virtual, cloud andphysical servers. Through deep integration with configurationmanagement, infrastructure services and PXE and Image-based unattendedinstallations, Foreman manages every stage of the lifecycle of yourservers. Foreman provides comprehensive, auditable interactionfacilities including a web frontend and robust, RESTful API.

Speaker: Gilad Chaplik

For several years now, the oVirt project is leveraging KVM and relevanttechnologies (ksm, etc) in data center virtualizations. Being a matureand feature reach, oVirt takes another step forward with introducing aPluggable Scheduling API. This presentation will review recent oVirtimprovements in the areas of VM scheduling. The first part will discussabout the architecture of the new scheduler. On the second part we willshow samples of VM scheduling plug-ins, and integrate it to a livesetup.

Speaker: Hans de Goede

Xorg (the X-server) is a big and complex beast. Currently it runs asroot as it needs root rights for various reasons. But with the latestsystemd-logind all necessary infrastructure is in place to allow theserver to run as a normal user and use systemd-logind to do input andgraphics device management.

This talk looks at the work being done to leverage this newinfrastructure to run Xorg without root rights.

Speaker: Langdon White

For many years the Linux distro concept has been about “inclusion ofapplications” sometimes at the detriment to co-habitating applicationsand the stability of the core OS. Much discussion has been made over theyears about JEOS, embedded Linux, custom distros, applicance building,etc, but not a lot of discussion about how applications could bedelivered such that they were more readily able to co-habitate. On arelated note, open source applications (because distros are so“inclusive”) are put through significant scrutiny around their designand deployment related to their integration with the core OS that may ormay not make sense. The scrutiny is certainly more intense thanproprietary software is required to undergo. This talk proposes anew’ish technology, Software Collections, as a method for resolution.

Speaker: Karanbir Singh

CentOS has cemented a reputation as the “community enterprise operatingsystem” - one that provides a reliable rebuild, but is not known forinnovation in its own right. With the news that Red Hat and CentOS arejoining forces, this is going to change. Here’s how CentOS is planningto change, and how other distros can learn from our next phase.

Speaker: Stephen Gallagher

As you may or may not be aware, Fedora is transitioning from its classic“one-size-fits-all” approach to one where we intend to target threespecific user types with individual products: Fedora Workstation, FedoraServer and Fedora Cloud. Gathering Fedora contributors at FOSDEM to workon the logistics around this change in direction would be a valuableopportunity.

Speaker: Christophe Fergeau, Pau Garcia i Quiles, Philippe Caseiro, Jerome Leclanche, Didier Roche

Presentation of the Desktops DevRoom by its Organization Team &Technical Committee: Christophe Fergeau (Gnome), Pau Garcia i Quiles(KDE), Didier Roche (Unity), Philippe Caseiro (Englightenment) andJérome Leclanche (LXDE)

Speaker: Lennart Poettering

With kdbus we move the D-Bus IPC system into the Linux kernel, toimprove performance and functionality, while keeping compatibility.

Speaker: Marek Hulán

In this talk I’d like to show a live demo covering status of Foreman andChef integration and try to answer the question “where do we want toget”? Also I could sum up what’s needed to add similar support forconfig management tools of your will.

Speaker: Tom Callaway

Recently, Cisco released a software component for H264 video supportunder the BSD license. This would normally be considered Free Software,except for the fact that this standard is considered to be patented(where software patents apply). They are covering the MPEG LA licensingfees, but only for pre-built binaries, not for the corresponding source.I will discuss this issue, and why the resulting software is not onlynon-free, but does not solve any issues for distributors with a focus onFreedom (such as Fedora).

Speaker: Stefano Zacchiroli, Karen Sandler, Christopher Webber, Eileen Evans, Tom Callaway, Chris Aniszczyk

Five of our speakers from the Legal and Policy Issues devroom haveagreed to participate in a governance best practices roundtable. Thesepractices may touch on contribution policy, review boards, policymanuals, licensing tools, trademark guidelines, etc. Questions will beasked of the panelists to start the roundtable and the audience willalso be encouraged to participate in order to have more interaction withthe panelists. Karen Sandler will be the moderator.

Speaker: Eike Rathke

ODF 1.2, additionally to the fo:language and fo:country attributes,introduced fo:script and *:rfc-language-tag attributes to allow for thefull range of BCP 47 language tags. This talk will give a brief overviewwhat it means to applications and how LibreOffice implemented it, andthe consequences it may have for extension developers.

Speaker: Stephan Bergmann

Advances in C++ have gathered momentum with C++11 and forthcoming C++14,and compiler writers busy to keep up. However, for reasons ofcross-platform, cross-compiler, aging baselines, etc., the LibreOfficecode base is still mostly stuck with C++98. We will discuss how toovercome blockers in adoption of modern C++ features and in what waysLibreOffice would benefit from them.

Speaker: Geoffrey De Smet

Some scientific research problems inherently suffer from an NP-completeproblem. This session will explain several metaheuristic algorithms,which can handle such problems in reasonable time.

This session will also do lightning introduction of OptaPlanner, an opensource apache licensed Java library, which implements those algorithms.

Specifically, these algorithms will be explained:

• First Fit
• First Fit Decreasing
• Hill Climbing
• Tabu Search
• Simulated Annealing
• Late Acceptance

Speaker: Soumya Deb

Starting with pdf.js, spiraling around shumway & zipfile.js, we’llexplore what JS is already capable of, that never seemed practical. Thenwe’ll go on to explain the WebAPIs to bring the “native” right in thebrowser (with a tinge of FxOS - so that it’s not up in the air, it’salready there - in fact it’s so-last-fosdem actually). Then we’ll talkabout the (near)future, and how broadway.js, asm.js (Emscripten, LLVM)et al. are going the change the web - for good!

Speaker: Andrew Haley, Andrew Dinn

Red Hat’s project of porting OpenJDK to run on ARM’s new 64-bitarchitecture began about 18 months ago. This talk will describe the workwe have performed over the last year, explaining how we went aboutimplementing the client and server JIT compilers. In particular, we willgive details and examples of how we have tuned the server compiler togenerate code that has been optimized to make use of the AArch64instruction set.

Speaker: Roman Kennke

Current garbage collectors for OpenJDK all need to stop the applicationperiodically to perform garbage collection tasks. This is a scalabilitybottleneck because those pause times are dependend on heap size.Shenandoah is a new garbage collector for OpenJDK, currently developedby Red Hat, that aims to reduce GC pause times to a minimum byimplementing marking and object evacuation to run concurrently withapplication threads, and utilizing parallel garbage collection threads.

Speaker: Charles Nutter

Write once run anywhere is both a blessing and a curse. For years, theWORA promise has ensured a consistent, compile-free experience for JVMusers. Unfortunately, sometimes pure-Java libraries just can’t do whatdevelopers need done. Sometimes, you just have to go full native. TheJava Native Runtime is a core library and suite of support libraries forbinding and calling out to native code. I’ll show how JNR is designed,compare code and performance with alternative approaches, and talk aboutwhy the JDK needs a standard FFI (foreign function interface) in Java 9.

Speaker: Mario Torre

Thermostat is an awesome platform with the focus on Java VirtualMachines monitoring. This short presentation will discuss what has beendone in the past two years until the release of Thermostat 1.0 and itscurrent features and will disclose some spoilers on the future ofThermostat.

Speaker: Mark Reinhold, Andrew Haley, Georges Saab, Doug Lea

An open Q&A session with members of the OpenJDK Governing Board.

Speaker: Pavel Zhukov

This presentation explains and demonstrates how the Fedora Linuxdistribution can be used for developing in the Ada language, tells aboutavailable tools and frameworks.

Speaker: Ric Wheeler

Persistent memory parts have roughly the same capacity, speed and costas current DRAM, but do not lose state when the power goes out. Some ofthese parts are on the market today, more will be coming out over thenext few years. The Linux IO and File System stack is already challengedin handling existing SSD devices at hundreds of thousands of IO’s persecond and these devices will be able to sustain an order of magnitudemore IOP’s.

This talk will give an overview of what is being proposed in standardsbodies and the Linux based solutions being proposed that will help ustake full advantage of these new parts.

Speaker: Nuno Diegues, Torvald Riegel

This talk will present Transactional Memory, a programming abstractionfor managing concurrency, both in multi-threaded programs running onmulti-core processors as well as in distributed cloud infra-structures.

Speaker: Richard Jones

Libguestfs (http://libguestfs.org) is a library and set of tools forsecurely and automatically performing many operations on disk images,from creating them to finding out what is in them.

In this 15 minute lightning talk, Richard Jones will talk about a few ofthe latest features, including the ability to churn out a new guestevery 60 seconds using virt-builder, how to really compress disk imagesusing virt-sparsify and xz. and “sysprepping” using virt-sysprep.

Speaker: Dave Neary

This talk will be a whirlwind tour around why you might want to use bothoVirt and OpenStack in your infrastructure, how you can use Gluster as acommon shared storage back-end for the whole thing, and what valueOpenShift adds into the mix. A full open source cloud on commodityhardware, in a couple of hours, covering virt, storage, IaaS and PaaS.

Speaker: Omer Frenkel

Getting cross-platform: bringing virtualization management to the PPCworld oVirt, the open source data center virtualization managementsolution is expanding the x86_64 architecture support to PowerPChardware, among them IBM Power processor based hosts.

Entering PPC world raises challenges of managing mixed arch data-centerswith the need to distinguish PPC from x86_64 VMs. In this session we’llcover the main differences in virtualization needs, how we enabled thesupport for that in our management engine(a.k.a ovirt-engine)

PPC virtualization main differences is in virtualized devices supports.Built around x86_64, ovirt-engine needed to expand its cluster and VMsprovisioning to fit multi architecture, and to keep migration supportand proper VM configuration solid for the supported operating systems.Since some parts are work in progress (e.g live migration) for the PPCVDSM, the engine has to also manage features which are not supported forPPC while is supported for x86_64

To cope with a change that span various parts and components arepository for Operating Systems and they’re hipervisor’s demands wascreated along with strategy functions around the code to ensure correctprovisions and life cycle related VM functions.

The session will cover what is now supported (hypervisor and VM OSs),how we provision PPC clusters and VMs and a little deeper dive toexplore the mechanisms which enabled this multi-platform support alongwith some code examples (yes, we’re going to see some code!)

Target audience Whoever is interested in data-center virtualization ingeneral and ovirt-engine specifically, and this latest PPC supporteffort it brings.

Speaker: Mike Kolesnik

As the prominent open-source data center virtualization solution, oVirthas many features that help you virtualize data center and cloudofferings. Sometimes a feature might be needed to extends oVirt’scapabilities, but even though oVirt is open source, you might want toprovide a quick and dirty solution..

Mike Kolesnik from Red Hat will show you how you can extends oVirt’scapabilities with ease throughout the oVirt stack - UI, engine and host.

Developers are welcome to join us in this session to learn how you canleverage oVirt to suit your virtualization needs.

Speaker: Doron Fediuck

For several years now, oVirt manages Virtual Machines. Then came thequestion- can you run oVirt inside a VM, which in turn will be managedby the hosted oVirt? In this session we’ll understand the intricacies ofan egg hosting it’s parent chicken. We’ll cover the various aspectsstarting with installation, going through standard operations and endingwith high-availability for the hosted engine. Participants will be ableto get insights of this unique setup, which will save them a physicalserver (or even two), while allowing standard flows to run the same waythey did in the past years.

Speaker: Assaf Muller

Join me for an architectural, developer oriented overview of (GRE andVXLAN) tunnels in OpenStack Networking.

In the virtualization environment virtual machines are hosted onhypervisors. These VMs then obtain network connectivity via softwareswitches run in the same hypervisors. Data centers that provideinfrastructure as a service have (hopefully) multiple customers (Or‘tenants’). As you can imagine we don’t want tenants’ VMs interactingwith one another.

VLANs are a natural approach to achieve tenant segregation. However, howdo we maintain scalability with a growing number of hypervisors and VMs,when the administrator has to constantly configure the hardware switchesmanually? Is there another way?

We all use VPNs to connect to our office resources remotely, or toconnect two office sites into one seamless network. VPNs are essentiallyencrypted tunnels, but what are tunnels?

Tunnels allow us to wrap packets inside more packets. In our context -VM traffic in exterior IP packets. That way, to the intermediatenetworking hardware, it looks like traffic between the hypervisors.Since the hypervisors should already be able to talk to each other, thismakes VM connectivity a breeze!

Let’s explore how tunnels are used in the cloud as a means to achieve anoverlay network. What is an overlay network? How does traffic flowbetween virtual machines on the same hypervisor, and on differenthypervisors? What are the similarities between a layer 2 learning switchand tunnel logic in OpenStack? How does Open vSwitch fit in? Is there acost to using tunnels?

This talk will be useful to developers interested in learning about newnetworking concepts - Minimal background knowledge will be assumed.

Speaker: Mike Kolesnik

As the prominent open-source data center virtualization solution, oVirtrelies on a powerful and easy approach to configuring a data center’snetwork. By leveraging the advanced network capabilities offered byOpenStack Networking, oVirt’s maintainers aim to bring this field evenfurther, allowing data center administrators to use advanced networkingcapabilities while maintaining the simplicity of oVirt’s networkmanagement approach.

Developers & Users are welcome to join us in this session, and todiscover how oVirt currently leverages OpenStack Networking, and see theroad-map to future network virtualization in the Data Center, all usingopen source enterprise-grade software.

Speaker: Nicolas Dandrimont, Pierre-Yves Chibon

Fedmsg, the federated message bus, is a distributed system allowing bitsof a project’s infrastructure to publish events. This lightweightframework provides a central place to watch the life of a project, andallows anyone to listen in and trigger actions when an event isreceived.

Initially built by the Fedora Infrastructure team to uniformisecommunication between its services, the Fedmsg framework has now beenadopted by Debian and even outside the realm of Linux distributions tobroadcast events. All those deployments are using the same framework,and the same tools, fostering the ever-needed interoperability andcollaboration between free software and open data projects.

In this talk, we will present how fedmsg works, how it was deployed inthe Fedora and Debian infrastructures, and some of the technical choicesthat were made during those deployments. We will also present theapplications being developed around the bus, as well as some ideas forfuture ones.

Speaker: Andreas Schneider

Testing network applications correctly is hard. The talk will show howto create a fully isolated network environment for client and servertesting on a single host, complete with synthetic account information,hostname resolution and privilege separation.

Speaker: Vadim Rutkovsky

We are a group of engineers from Red Hat’s Desktop QE team and we wouldlike to discuss standalone application testing on Linux. During thisworkshop we’d like to show existing workflow of application testing,discuss testing tools and overall influence of quality engineers on opensource software development process.

Speaker: Greg Sutcliffe

A look at how Foreman uses automation internally to handle testing andrelease management.

Speaker: Roland Grunberg

For developers it can often be a bit of a learning curve to learn theproper use of a new tool. For certain development tools the entrybarrier can be quite high and can often discourage users. The LinuxTools Project aims to improve the state of C/C++ development on theEclipse IDE by integrating popular tools, such as Valgrind.

This talk is aimed at people of varying experience with the Valgrindtool who have never used it within the Eclipse IDE.

Speaker: Miroslav Franc

Valgrind is a tool that can be used for testing but also needs to betested itself as any other piece of software. This talk will focus ontesting done before releasing new Valgrind RPM in Red Hat EnterpriseLinux.

Overview: - types of tests I do - what they are trying to achieve - whatbugs I often hit - ideas I have for improvement - Q/A + Discussion(perhaps how to work closely with upstream)

Speaker: Pedro Alves

In this talk I’ll present an overview of the current state of severalGDB projects of interest to Valgrind developers looking at GDB/Valgrindintegration, including the current state of GDBserver in GDB, where weare on local vs remote feature parity, all-stop vs non-stop modes,multi-process and multi-target projects, reverse debugging and possiblyothers.

• current state of GDBserver in GDB
• where we are on local vs remote feature parity
• all-stop vs non-stop modes
• multi-process debugging
• multi-target projects
• reverse debugging

Speaker: Tom Tromey

Given the current state of Valgrind and GDB how can we make things evenbetter and smoother. Put some Valgrind and GDB hackers in the same roomand let them discuss the technical details needed on each side. Come andhelp us brainstorm some crazy and fun ways to make the Valgrind/GDBcombo even cooler and more powerful.

Speaker: Mark Wielaard

Valgrind developers and users are encouraged to participate either bysubmitting ideas/suggestions or by joining the discussion. And of courseby kindly (or bitterly:) complain about bugs you find important that arestill Not YET solved since that many years !?@^!!!