Loading…
Vault 2017 has ended
Be sure to join us at Vault 2017 and register here

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

Tuesday, March 21
 

2:00pm EDT

Pre-Registration
Tuesday March 21, 2017 2:00pm - 5:30pm EDT
Courtyard Reception
 
Wednesday, March 22
 

8:00am EDT

Breakfast
Wednesday March 22, 2017 8:00am - 9:00am EDT
President's Ballroom A+B

8:00am EDT

Sponsor Showcase
Wednesday March 22, 2017 8:00am - 12:15pm EDT
President's Ballroom A+B

8:00am EDT

Registration
Wednesday March 22, 2017 8:00am - 6:00pm EDT
Courtyard Reception

9:00am EDT

Keynote: A Perspective on the Future of Storage - Garth Gibson, Professor of Computer Science, Carnegie Mellon University (CMU)
Speakers
avatar for GARTH GIBSON

GARTH GIBSON

Professor & Assoc. Dean, CARNEGIE MELLON UNIVERSITY
Gibson is a full professor of Computer Science at Carnegie Mellon University (CMU), with a courtesy appointment in the Department of Electrical and Computer Engineering, and the Associate Dean for Master’s Programs in the School of Computer Science at CMU. His PhD (1991) and MS... Read More →


Wednesday March 22, 2017 9:00am - 9:20am EDT
President's Ballroom C+D

9:25am EDT

Keynote: The Forever Kernel - Chris Mason, Software Engineer, Facebook

Join Facebook Kernel Engineer Chris Mason as he reads excerpts from the Kernel Team's new book: The Forever Kernel


Speakers
avatar for Chris Mason

Chris Mason

Software Engineer, Facebook
Chris is a Software Engineer on the kernel team at Facebook, and the maintainer of the Btrfs filesystem. He has been working full time on the kernel for over 16 years, and lives in Rochester New York.


Wednesday March 22, 2017 9:25am - 9:45am EDT
President's Ballroom C+D

9:45am EDT

LSF-MM Summit Readouts from Chairs Josef Bacik, Facebook; Martin Petersen, Oracle and Rik van Riel, Red Hat
The three track leaders from the recent LSF/MM Summit will present the outcomes from the summit and some indications of the future directions the Linux ecosystem is taking.
  • Martin Petersen, Oracle - IO Track Leader
  • Josef Bacik, Facebook - FS Track Leader
  • Rik van Riel, SUSE Labs - MM Track Leader 

Moderators
avatar for Jeff Layton

Jeff Layton

Principal Software Engineer, Red Hat
Jeff Layton is a long time Linux kernel developer specializing in network file systems. He has made significant contributions to the kernel's NFS client and server, the CIFS client and the kernel's VFS layer. Recently, he has taken an interest in Ceph, in particular as a backend for... Read More →

Speakers
JB

Josef Bacik

Facebook
avatar for Martin K. Petersen

Martin K. Petersen

Architect, Oracle
Martin K. Petersen has contributed to the Linux Kernel since the early nineties. As Linux Storage Architect at Oracle he focuses on future I/O and storage technologies. Martin also participates in various industry standardization efforts and co-maintains the Linux SCSI stack.
RV

Rik Van Riel

Principal Software Engineer, Red Hat
Rik van Riel is a principal software engineer at Red Hat, and a long term contributor to the Linux kernel. He has contributed to the memory management subsystem, the scheduler, and various components related to virtualization. Rik is active in community projects like kernelnewbies.org... Read More →


Wednesday March 22, 2017 9:45am - 10:25am EDT
President's Ballroom C+D

10:25am EDT

Coffee Break
Wednesday March 22, 2017 10:25am - 11:00am EDT
President's Ballroom A+B

11:00am EDT

Evolving Ext4 for Shingled Disks - Abutalib Aghayev, CMU & Theodore Ts'o, Google
Drive-Managed Shingled Magentic Recording (SMR) disks offer a higher capacity alternative to traditional disk drives. However, non-sequential workloads can show bi-modal behaviour. After a short period of high performance they enter a continuous period of low performance. We were able to make a small change (600 LOC) to ext4 that significantly improves the throughput in both modes, resulting in 2-13x improvements on metadata-heavy workloads, and 1.7-4.9x improvements on a file server benchmark. The changes also resulted in performance improvements on conventional disk drives.

Speakers
AA

Abutalib Aghayev

Graduate Student, Carnegie Mellon University
Abutalib Aghayev is a PhD student in the Computer Science department at Carnegie Mellon University.
TT

Theodore Ts'o

Staff Programmer, Google
Theodore Ts'o is the first North American Linux Kernel Developer, and started working with Linux in September, 1991. He previously served as CTO for the Linux Foundation, and is currently employed at Google. Theodore is a Debian Developer, and is the maintainer of the ext4 file system... Read More →


Wednesday March 22, 2017 11:00am - 11:50am EDT
Thomas Paine AB
  Filesystems

11:00am EDT

Challenges in Management Services for Distributed Storage - Mrugesh Karnik, Red Hat
Software defined storage is an important part of a datacentre's infrastructure. Providing a well designed resilient management service that can be integrated across existing management platforms adds value to large infrastructure deployments.

Tendrl is an actively developed storage management system that provides a unified infrastructure and framework. This framework allows diverse management flows to be built and deployed dynamically for different storage systems. The current release of Tendrl supports Ceph and Gluster and provides management and monitoring capabilities for these systems.

In this presentation Mrugesh will be discussing the Tendrl architecture and the design choices that have shaped it. Mrugesh will also provide a look at how the core Tendrl framework enables dynamic support for different storage systems.

Speakers
MK

Mrugesh Karnik

Principal Software Engineer, Red hat
Mrugesh Karnik is the Project Technical Lead for Tendrl and is a Software Engineer at Red Hat. Mrugesh has previously conducted workshops around system provisioning and monitoring technologies, large system administration and is a coach for software developers writing toolkits for... Read More →


Wednesday March 22, 2017 11:00am - 11:50am EDT
Paul Revere C
  Management

11:00am EDT

Running GlusterFS at Facebook Scale: Performance, Debugging, Tooling - David Hasson, Shreyas Siravara & Kevin Vigor, Facebook
Facebook runs several storage platforms at scale, one of them being GlusterFS - a distributed POSIX filesystem that runs well on commodity hardware. These systems can be challenging to run reliably at scale. As the amount of data, I/O rate, and number of clients go up, both problem diagnosis and system maintenance activities become increasingly difficult. In this talk, we'll talk about the tools and techniques that we use to administer and debug large distributed filesystems and show how others can leverage these techniques in their own environments.

Speakers
DH

David Hasson

Production Engineer, Facebook Inc.
David works on the POSIX Storage team at Facebook. The team maintains GlusterFS at Facebook and also runs some proprietary storage technologies at scale. Previous to Facebook, David has worked at EMC/Isilon and in research computing at UCLA.
KV

Kevin Vigor

Software Engineer, Facebook
Kevin is currently employed at Facebook working on the Gluster distributed filesystem and surrounding ecosystem. He was previously employed at Fusion-IO working on high-performance (and sadly proprietary) linux block storage drivers. Significant experience presenting to engineering... Read More →


Wednesday March 22, 2017 11:00am - 11:50am EDT
Paul Revere AB

11:00am EDT

Open-Channel SSDs: Then. Now. And Beyond - Matias Bjørling, CNEX Labs
The Open-Channel SSD kernel subsystem is maturing, the first block-device FTL (pblk) is ready for upstream kernel, and the liblightnvm user-space library is ready for use. This talk will cover an introduction to Open-Channel SSDs, an update on current progress, and what the community is working on. We show how pblk is initialized, and show its performance on an Open-Channel SSD. At last, we provide an overview of the LightNVM architecture, and how one can develop for Open-Channel SSDs.

Speakers
MB

Matias Bjørling

LightNVM Architect, CNEX Labs
Matias Bjørling is the kernel maintainer of the LightNVM subsystem and leads the Open-Channel SSD efforts at CNEX Labs. Before joining the industry, he obtained a Ph.D. in operating systems, and non-volatile storage by doing performance characterization of flash-based SSDs, working... Read More →


Wednesday March 22, 2017 11:00am - 11:50am EDT
William Dawes AB
  Solid State

12:00pm EDT

Stratis: A New Approach to Local Storage Management - Andy Grover, Red Hat
Linux's existing choices for volume-managing filesystems (VMFs) have been limited to two choices: either ZFS-on-Linux, which has licensing issues; or Btrfs, still seeking wider adoption. Stratis aims to deliver equivalent features and ease of use as a VMF, but with a hybrid, layered approach that builds on Linux's strengths to deliver a stable solution within a very short time window.

Andy Grover, Stratis team lead and architect, will talk about the advantages and trade-offs in Stratis's design, and some controversial development decisions. He will also discuss its current implementation state, and share plans for development goals leading up to a stable 1.0 release, and beyond.

Speakers
AG

Andy Grover

Principal Software Engineer, Red Hat
Andy Grover is a Principal Software Engineer at Red Hat, working to improve Linux’s block storage capabilities. His work encompasses both changes to the kernel itself as well as low-level management tools. Previous work areas include networking and ACPI. He has previously spoken... Read More →


Wednesday March 22, 2017 12:00pm - 12:50pm EDT
Thomas Paine AB
  Filesystems
  • Experience Level Any

12:00pm EDT

DeepSea: The Salt solution for Ceph - Eric Jackson, SUSE
Are you interested in creating Ceph clusters quickly? What about growing, shrinking and reallocating monitors, storage nodes or gateways without having to recall all the specific Ceph commands?

DeepSea is a collection of Salt files with two simple conventions: Stages and Roles. The primary goals are reasonable working defaults, complete customization and quick debugging. The most important goal is usability. The Stage and Role abstractions are simple, boring and lend themselves well to automation.

Come learn how to deploy and manage a Ceph cluster and its services with Salt.

Speakers
EJ

Eric Jackson

Senior Software Developer Distributed Storage, SUSE
I am part of the storage team at SUSE. My primary project is DeepSea https://github.com/SUSE/DeepSea. I also maintain lrbd https://github.com/SUSE/lrbd. I have given three talks: one at SUSECon 2016 and two at SUSECon 2017.


Wednesday March 22, 2017 12:00pm - 12:50pm EDT
Paul Revere C
  Management
  • Experience Level Any

12:00pm EDT

Large-scale Stability and Performance of the Ceph File System - Patrick Donnelly, Red Hat
The Ceph file system (CephFS) is the POSIX-compatible distributed file system
running on top of Ceph's powerful and stable object store. This presentation
details recent performance measurements and stability testing of CephFS in
large-scale environments via virtual private servers in the cloud. Particular
emphasis is placed on exploring the benefits and stability of using multiple
metadata servers to distribute metadata load, a CephFS feature slated for
stability in the upcoming Luminous release. This talk will share the newer
methods used for rigorously testing CephFS in larger cloud environments by the
CephFS team and how we are measuring performance.

Speakers
PD

Patrick Donnelly

Senior Software Engineer, Red Hat, Inc.
Patrick Donnelly is a senior software engineer at Red Hat, Inc. currently leading the global development team working on the open-source Ceph distributed file system. Patrick has been a speaker at several events presenting recent work on Ceph, including Cephalocon APAC, various Openstack... Read More →


Wednesday March 22, 2017 12:00pm - 12:50pm EDT
Paul Revere AB
  Software Defined
  • Experience Level Any

12:00pm EDT

Improving Block Discard Support throughout the Linux Storage Stack - Christoph Hellwig
Flash based storage supports the concept of discarding data in blocks without actually overwriting, a concept that helps with the internal wear level and data placement algorithms. Linux has supported this concept, which has different names in different storage protocols (trim, unmap, deallocate), for a long time. But the concept of "online" or live discards that notify the device instantly after the deletion of data in the file system has only seen limited traction in Linux, mostly due to the severe performance degradation caused by it.

This talks explains optimizations to the file system and block layer to allow better batching and asynchronous execution of discard requests, as well how the file system block allocator can better be aware of ongoing discards.

It will also explore how discarding of data overlaps with fast zeroing operations, and why it really shouldn't at the interface level.

Speakers
CH

Christoph Hellwig

Christoph Hellwig has been working on Linux Storage and File system projects for 15 years. He works all the way up and down the Storage and File system stack, and runs a business focused on Linux Storage architecture and training.


Wednesday March 22, 2017 12:00pm - 12:50pm EDT
William Dawes AB
  Solid State

12:50pm EDT

Lunch (Attendees on Own)
Wednesday March 22, 2017 12:50pm - 2:50pm EDT
TBA

2:30pm EDT

Sponsor Showcase
Wednesday March 22, 2017 2:30pm - 4:00pm EDT
President's Ballroom A+B

2:50pm EDT

The Extent of GFS2 - Steven Whitehouse, Red Hat
The GFS2 cluster filesystem has proved itself as a robust and reliable filesystem for cluster use cases. Most of the recent development has been focussed on performance improvments, and improving the ease of use and deployment. This talk will cover recent developments in GFS2, but also discuss future plans too. When GFS was originally designed, it took a lot of inspiration from ext2/3 which resulted in bitmap based resource groups, and an equal height pointer tree for inode metadata. The question arises as to whether and how GFS2 might integrate support for extents. The talk will present the issues involved and the latest thoughts of the development team on this topic.

Speakers
SW

Steven Whitehouse

Senior Manager, RHEL Filesystems, Red Hat
Steven Whitehouse currently manages the RHEL Filesystems team at Red Hat. His introduction to Linux kernel development came in 1993 when he wrote a small patch for AX.25, he is also the previous maintainer of Linux DECnet and the GFS2 Filesystem. Steven has spoken at a number of conferences... Read More →


Wednesday March 22, 2017 2:50pm - 3:40pm EDT
Thomas Paine AB
  Filesystems

2:50pm EDT

Refreshing 100+ TB Per Day Across Data Centers with Voldemort - Felix GV, LinkedIn
Voldemort is an open-source distributed key-value store with the capability to bulk load data from Hadoop.

At LinkedIn, Voldemort is deployed in a geographically distributed environment across hundreds of servers. Each day, 100+ TB of data need to be refreshed, while continuing to serve sub-millisecond read requests.

This talk will describe the Voldemort architecture, with deep dives on the Read-Only storage format, the compression strategies and the bandwidth minimization schemes that it employs. The design choices and tradeoffs of the system will be explained.

Speakers
FG

Felix GV

Data Infrastructure Engineer, LinkedIn
Felix GV is an engineer on LinkedIn's Data Infrastructure. He works on the Voldemort and Venice projects, supporting use cases that need to refresh 100+ TB per day of data in a geographically distributed fashion, while providing low latency random access to local applications. His... Read More →


Wednesday March 22, 2017 2:50pm - 3:40pm EDT
Paul Revere C
  Management

2:50pm EDT

Geo-Replication and Disaster Recovery for Cloud Object Storage with Ceph Rados Gateway - Orit Wasserman, Red Hat
Ceph is a highly available distributed software defined storage, providing object, key/value and file-system interfaces. Ceph RGW (Rados Gateway) provides HTTP REST API that is S3 and openstack swift compatible. Many users need storage systems that can span multiple data centers and geographies for disaster recovery and for better time response in remote locations. This talk will give a brief Ceph architecture overview and then focus on the design and the new implementation of asynchronous Geo-Replication and disaster recovery features in Ceph Rados Gateway. We will also describe its configuration and usage.

Speakers
avatar for Orit Wasserman

Orit Wasserman

Senior Principal Software Engineer, Red Hat
Orit is a senior principal software engineer at Red Hat, focusing on Container and multi cloud storage. She was a principal architect at Lightbits labs working on NVMe/TCP software-defined storage. At Red Hat, she worked on Ceph object storage (Ceph Rados Gateway), a highly available... Read More →


Wednesday March 22, 2017 2:50pm - 3:40pm EDT
Paul Revere AB

2:50pm EDT

The Unwritten Contract of Solid State Drives - Jun He
Jun He has performed a detailed vertical analysis of application performance atop a range of modern file systems and SSD FTLs. He has formalized the "unwritten contract" that clients of SSDs must follow to obtain high performance, and conduct the analysis to uncover application and file system designs that violate the contract. The analysis, which utilizes a highly detailed SSD simulation underneath traces taken from real workloads and file systems, provides insight into how to better construct applications, file systems, and FTLs to realize robust and sustainable performance. In this presentation, Jun He will present the methodology of the vertical analysis, as well as the insights learned from applications (e.g., LevedlDB, RockDB, SQLite-WAL, SQLite-RollBack, and Varmail), file systems (e.g. ext4, F2FS, and XFS) and SSD simulations.

Speakers
avatar for Jun He

Jun He

PhD student, University of Wisconsin-Madison
Jun He is a fourth-year PhD student at Department of Computer Sciences, University of Wisconsin-Madison. He works with Prof. Andrea C. Arpaci-Dusseau and Prof. Remzi H. Arpaci-Dusseau. He has been studying and designing tools to help system developers to better understand interactions... Read More →


Wednesday March 22, 2017 2:50pm - 3:40pm EDT
William Dawes AB
  Solid State

3:40pm EDT

Coffee Break
Wednesday March 22, 2017 3:40pm - 4:00pm EDT
President's Ballroom A+B

4:00pm EDT

Panel Discussion: Present and Future of Persistent Memory Support in Linux - Moderated by Dan Williams, Intel
The kernel developers on this panel are at the forefront of persistent memory enabling in Linux. Persistent memory changes fundamental assumptions about storage in the kernel and this has led to healthy debates about how to evolve the kernel into a persistent memory future. The discussion will range from the infrastructure that the kernel has gained to date and look forward to the capabilities that are on the horizon.

Moderators
avatar for Dan Williams

Dan Williams

Software Engineer, Intel
Dan is a Linux kernel developer in Intel Open Source Technology Center primarily focused on storage technologies. Most recently he has been involved in persistent memory enabling as a maintainer of the Linux kernel libnvdimm sub-system. He led the Persistent Memory micro-conference... Read More →

Speakers
CH

Christoph Hellwig

Christoph Hellwig has been working on Linux Storage and File system projects for 15 years. He works all the way up and down the Storage and File system stack, and runs a business focused on Linux Storage architecture and training.
avatar for Jan Kara

Jan Kara

Kernel engineer, SUSE
Jan Kara is doing Linux kernel hacking in file systems area over 15 years. He is the maintainer of udf file system, and quota subsystem. He is also working on ext4 and other filesystems, writeback logic, notification framework, and other miscellaneous stuff. Currently he is working... Read More →
LK

Linda Knippers

Distinguished Technologist, Hewlett Packard Enterprise
JM

Jeff Moyer

Senior Principal Software Engineer, Red Hat, Inc.
avatar for Visha Verma

Visha Verma

Software Engineer, Intel
Vishal works in Intel's OTC, and is part of the team enabling NVDIMMs on Linux. Talk to him about NVDIMMs, DAX, Sector Atomicity, Media errors on NVDIMMs and how they are handled.
avatar for Ross Zwisler

Ross Zwisler

Linux Kernel Developer, Intel
Ross is a Linux kernel developer currently focused on enabling persistent memory. His primary contributions have been to the PMEM and ND BLK persistent memory block drivers and to the DAX (direct access) filesystem feature.


Wednesday March 22, 2017 4:00pm - 4:50pm EDT
Thomas Paine AB
  Filesystems
  • Experience Level Any

4:00pm EDT

DataCenter.ai - Analytics for Modern Data Centers, and A Storage Case Study for Disk Failure Prediction - Yuming Ma, Cisco
Data Center management and operation have been a challenging element for IT infrastructure. We proposed a framework to address the key problems of workload prediction & balance; resource overhead & waste reduction; component/disk failure and performance degradation by combining data analytics and machine learning technologies. We will illustrate the analytics intelligence with a case study of disk failure prediction which has been one of the fundamental problem in distributed storage systems. The disk failure prediction uses SMART data collected from each disk which are then crunched through an analytics engine empowered by machine learning algorithms to predict disk failure with 90% of false positive rate and two weeks of lead time. We also believe this framework will serve as the stepping stone for data center modernization, automation, and orchestration

Speakers
YM

Yuming Ma

Architect, Cisco


Wednesday March 22, 2017 4:00pm - 4:50pm EDT
Paul Revere C
  Management

4:00pm EDT

Improving Performance of Directory Operations in Gluster - Manoj Pillai & Raghavendra Gowdappa, Red Hat
Providing good performance for directory operations in a scale-out storage solution like gluster can be challenging. Gluster relies on a framework called Distributed Hash Table (DHT) for hashing and distributing files, which makes it necessary to aggregate contents of a logical gluster directory from multiple physical directories on the underlying constituent file systems, known as bricks, for operations like readdir.

In this talk, Manoj Pillai will present an extensive evaluation of the performance of gluster directory operations for real-life scenarios. Raghavendra Gowdappa and Poornima Gurusiddaiah will then describe the following approaches for improving directory performance, and assess their effectiveness: readdir-ahead (prefetching), parallel readdir, readdir-optimize (filter duplicate directories that result from gluster's directory creation approach) and directory readdirplus.

Speakers
RG

Raghavendra Gowdappa

Principal software Engineer, Redhat India Pvt Ltd
avatar for Manoj Pillai

Manoj Pillai

Senior Principal Software Engineer, Red Hat
Manoj Pillai is part of the Performance and Scale Engineering Group at Red Hat, with focus on storage performance. He has presented previously at Open Source Summit Europe, and also at other open source conferences such as Cephalocon, Vault, FOSDEM and Gluster Summit.


Wednesday March 22, 2017 4:00pm - 4:50pm EDT
Paul Revere AB

4:00pm EDT

I/O Latency Optimization with Polling - Damien Le Moal, Western Digital
I/O completion polling is a very effective technique for reducing I/O latency under certain workloads. Recent changes to the Linux kernel have greatly improved the efficiency of the I/O polling code in two major areas: (1) lower code overhead for small raw block device direct I/O accesses and (2) CPU load cost of I/O polling through "hybrid" polling.
This presentation will explain in detail these changes and show many performance evaluation results using fast NVMe SSDs to illustrate how effective polling can be compared to a regular interrupt based I/O completion scheme. The gains obtained with recent optimizations will also be numerically detailed.

Speakers
avatar for Damien Le Moal

Damien Le Moal

Director, Western Digital
Damien Le Moal manages the System Software Group within Western Digital Research. Damien is a regular contributor to Linux kernel block I/O stack and file system support for the ZBC and ZAC standards (zoned block devices) and is a maintainer of the kernel dm-zoned device mapper and... Read More →


Wednesday March 22, 2017 4:00pm - 4:50pm EDT
William Dawes AB
  Solid State

5:00pm EDT

XFS Development AMA - Darrick Wong, XFS Maintainer
XFS has long been the scale-up scale-out filesystem first for SGI Irix and now Linux. Recent developments in XFS will soon enable users to perform lightweight file copies, deduplicate file data, perform online filesystem checking and repair, and take advantage of persistent memory.

In the first part of this session, the presenter will discuss the motivation behind the new features, sketch how they are implemented, and talk about how they affect everyday operations of a data store. Possible future development direction will also be presented, along with a summary of relevant LSFMM discussions.

In the second part of the session, the floor will be opened to the audience to ask questions about existing and upcoming features, make suggestions about what kinds of things they'd like to see from the Linux filesystem community over the next few years, and to share their own user stories.

Speakers
DW

Darrick Wong

XFS BOFH :-P, Oracle
I am the XFS maintainer. Previously I worked on ext4 and other parts of the storage stack. I have worked on Linux filesystems at Oracle, IBM, and Sun.


Wednesday March 22, 2017 5:00pm - 5:50pm EDT
Thomas Paine AB
  Filesystems

5:00pm EDT

Predicting Storage Failures with Machine Learning - Ahmed El-Shimi, Minima
Disk drives fail at an average annual rate of ~2%. Any system with Availability and Durability requirements must mitigate for such failures through a redundancy technique such as RAID, Erasure Coding, Replication or Backup.

With the wealth of monitoring data available nowadays and the ability to process the data in near real-time, can we predict such failures? How well can we do it? And how would that impact how we design and operate large distributed systems?

We examine and motivate predictive failure detection in the context of Availability, Rebuild Times and Recovery Objectives of large systems. We then train and evaluate multiple models achieving favorable accuracy (97.5%) to common datacenter practices. We demonstrate how we can tune our learners to achieve different Precision and Recall objectives thus improving Availability, Protection or Operational Efficiency.

Speakers
avatar for Ahmed El-Shimi

Ahmed El-Shimi

Founder, Minima
Ahmed El-Shimi has worked in Storage, Distributed Systems, and Cloud for over 15 years. He built technologies such as Deduplication, Automated Tiering, Hybrid Cloud Storage and Data Awareness. He is currently Co-Founder of Minima Inc. a Cloud Data Governance Startup. Prior he led... Read More →



Wednesday March 22, 2017 5:00pm - 5:50pm EDT
Paul Revere C
  Management

5:00pm EDT

Disaster Recovery and Ceph Block Storage: Introducing Multi-Site Mirroring - Jason Dillaman, Red Hat
Ceph provides a resilient and scalable software defined storage model. Ceph's block storage layer, RADOS block device (RBD), is widely used for virtual machine disks and is the most popular block storage used with OpenStack.

One of the most requested capabilities for Ceph has been the ability to continuously mirror RBD images to an offsite, disaster recovery location. Previously, disaster recovery with Ceph was a manual process, using differential snapshots and custom scripts. With multi-site mirroring, Ceph gains the ability to have near real-time consistency in separate geographical locations.

In this session, we’ll introduce the RBD journaling and mirroring features that provide this new ability. We’ll provide an overview of the architecture, explain the intended use cases, and show how to configure mirroring between Ceph clusters.

Speakers
JD

Jason Dillaman

Principal Software Engineer, Red Hat
Jason Dillaman is a principal software engineer at Red Hat, has been an active contributor to the Ceph project for over five years, and has been developing distributed systems for over 15 years. Jason currently acts as the project technical lead for the RADOS Block Device (RBD) team... Read More →


Wednesday March 22, 2017 5:00pm - 5:50pm EDT
Paul Revere AB

5:00pm EDT

Toward Robust Hot-pluggable PCIe Storage - Keith Busch, Intel
NVM Express has brought PCIe attached storage to the mainstream, and we're seeing this fast interface replacing solid state storage on slower, albeit more mature, buses. There is a strong user expectation on maintaining feature parity with those mature interfaces, and surprise hot plug is the one feature that trips up testing most often. This is not surprising as many software and hardware components that were previously not in the storage hot plug path need to come together to harmoniously work in any given situation. This presentation will explain the most recent software and hardware enhancements to handle corner case and error conditions, as well as remaining gaps and limitations.

Speakers
avatar for Keith Busch

Keith Busch

Software Engineer, Intel Corporation
Keith Busch is a software engineer working for Intel's Non-Volatile Solutions Group (NSG). He develops NVM products, writes updates and maintains the Linux drivers and user software, and represents Intel at committees and various industry conferences.


Wednesday March 22, 2017 5:00pm - 5:50pm EDT
William Dawes AB

6:00pm EDT

7:00pm EDT

Birds of a Feather Session - NFS - Steve Dickson, Red Hat
Speakers
SD

Steve Dickson

Consulting Software Engineer, Red Hat
Steve Dickson is a kernel Engineer at Red Hat. For years he as been the tech lead for the NFS team. He maintains a number of upstream packages and is responsible for the NFS offerings in both the RHEL and Fedora distros. In prior lives he worked on DEC's TruCluster product as well... Read More →


Wednesday March 22, 2017 7:00pm - 8:00pm EDT
Thomas Paine AB

7:00pm EDT

Birds of a Feather Session - The Global AFS Namespaces for Containers: A Secure Distributed Filesystem Approach to Securely Provide Persistence to Linux Containers

The AFS secure distributed file system provides a global namespace with access transparent paths to unstructured data and applications using the same file/directory paths without local administration or tooling.  For containers to take advantage of this, the AFS namespace must be available as a Kernel Level file system in standard Linux distributions.  This allows containers to have access, under the same global namespace, as any desktop/server/device within an organization. With key/credential injection by container orchestration tools and container process isolation, container images can be heterogeneously and securely deployed in your data center or any private/public cloud and have transparent access to persistent organizational storage.   The AFS namespace can also be used as a secure repository for container binaries.

This BOF follows the all-day AF_RXAPI and kAFS Hackathon.  In addition to discussing issues relating to using the Global AFS namespace for containers, we will discuss the current state and future of kAFS and the RX Linux Kernel implementations and how having kernel support for AFS and the AuriStor extensions to AFS positively affects security, support, distribution and performance.


Wednesday March 22, 2017 7:00pm - 8:30pm EDT
Crispus Attucks

7:00pm EDT

 
Thursday, March 23
 

8:00am EDT

Breakfast
Thursday March 23, 2017 8:00am - 9:00am EDT
President's Ballroom A+B

8:00am EDT

Sponsor Showcase
Thursday March 23, 2017 8:00am - 12:15pm EDT
President's Ballroom A+B

8:00am EDT

Registration
Thursday March 23, 2017 8:00am - 4:25pm EDT
Courtyard Reception

9:00am EDT

Keynote: Lessons Learned From NFS - Anna Schumaker, Engineer, NetApp

Anna will be talking about problems with the pace of NFS development and lessons learned while implementing the new features added in NFS v4.2.  Anna will also touch on a proposal for avoiding these issues and getting out new features faster in the future.

 


Speakers
avatar for Anna Schumaker

Anna Schumaker

NFS Client Maintainer, Netapp
Anna started working on Linux / NFS in 2010, and has been co-maintainer of the Linux NFS client since 2015. Recently, she has been working on implementing the new features in the NFS v4.2 protocol.


Thursday March 23, 2017 9:00am - 9:20am EDT
President's Ballroom C+D

9:25am EDT

Keynote: Alluxio (Formerly Tachyon): The Journey thus Far and the Road Ahead - Haoyuan Li, CEO, Alluxio
Alluxio (formerly Tachyon) is an open source memory-speed virtual distributed storage system. The Alluxio community is one of the fastest growing open source communities in big data history, with more than 400 developers from over 100 organizations around the world, and the Alluxio system has been deployed at a number of companies, including Alibaba, Baidu, Barclays, Intel, Huawei, and Qunar. In some of these deployments, Alluxio has been running in production for over a year, managing petabytes of data.

In the past year, the Alluxio project experienced significant improvement in performance and scalability and was extended with key new features including tiered storage, transparent naming, and unified namespace. At the same time, the Alluxio ecosystem has expanded to include support for more under storage systems and computation frameworks. In particular, Alluxio now supports a wide range of under storage systems, including Amazon S3, Google Cloud Storage, Gluster, Ceph, HDFS, NFS, and OpenStack Swift. These integrations make it possible for Alluxio to be leveraged in many different environments.

Haoyuan Li explores Alluxio’s goal of making its product accessible to an even wider set of users, through a focus on security, new language bindings, and further increased stability. Haoyuan also covers some new APIs Alluxio is working on to allow applications to access data more efficiently and manage data across different under storage systems.

Speakers
avatar for Haoyuan (H.Y.) Li

Haoyuan (H.Y.) Li

Founder and CEO, Alluxio
Haoyuan (H.Y.) Li is the Founder, Chairman, and CEO of Alluxio. He holds a Ph.D. in computer science from UC Berkeley’s AMPLab, where he created the Alluxio (formerly Tachyon) open source data orchestration system, and co-created Apache Spark Streaming as an Apache Spark founding... Read More →


Thursday March 23, 2017 9:25am - 9:45am EDT
President's Ballroom C+D

9:45am EDT

Coffee Break
Thursday March 23, 2017 9:45am - 10:25am EDT
President's Ballroom A+B

10:25am EDT

A Tail of Latency, IOPs, and IO Priorities - Adam Manzanares, Western Digital
Modern hard disk drives can achieve higher throughput for small IO accesses through queuing and scheduling optimizations within the device. Unfortunately, higher throughput for small IO comes at the cost of larger tail latencies. In this presentation, we will explore one option to improve tail latencies while minimizing the impact to throughput by leveraging iopriority information. I will discuss the recent patches that were merged in the Linux Kernel that enable SATA HDDs to leverage iopriority information. In addition, this presentation will cover how to use iopriority and provide some measured results.

Speakers
AM

Adam Manzanares

Research, Western Digital
I am a research engineer working for Western Digital focused on storage systems software. I am currently researching and developing systems software that is compatible with SMR HDDs, latency optimizations in devices and storage stacks, and distributed storage systems. I have given... Read More →


Thursday March 23, 2017 10:25am - 11:15am EDT
William Dawes AB
  Block

10:25am EDT

Persistent Storage for Containers with Gluster in Containers - Michael Adam, Red Hat
Containers are built to be ephemeral, stateless entities. But stateful applications running inside containers need a place to persistently store their data. The topic of persistent storage in container orchestration platforms like kubernetes is still relatively new and usually requires a lot of administrator activities and extra storage hardware.

This talk introduces gluster-kubernetes, a new project that deploys one or more GlusterFS clusters along with the volume management interface heketi as containers into kubernetes with a single command. This gives the administrator a robust and scalable persistent storage solution for their containers which is seamlessly integrated with kubernetes itself and allows to dynamically storage volumes of requested sizs for an application. The talk will feature a demo, and explain the roadmap of gluster-kubernetes.

Speakers
avatar for Michael Adam

Michael Adam

Principal Software Engineer, Red Hat
Michael Adam is an enthusiastic open source software developer, interested in all things about storage and containers. One of the main developers of Samba since more than a decade, Michael is an engineering manager at Red Hat, leading two worldwide teams: The Samba team for Gluster... Read More →


Thursday March 23, 2017 10:25am - 11:15am EDT
Paul Revere AB
  Containers

10:25am EDT

Campaign Storage - Gary Grider, Los Alamos National Laboratory
Computing sites need long-term retention of cool data often “data lakes” which focus on capacity but with non trivial bandwidth requirements. For many years, tape was the best economic solution but bandwidth and access needs have outstripped tape solutions. Disk can be more economically for this storage tier. The Cloud Community uses erasure based object stores to gain scalability and durability using commodity hardware. The Object Interface works for new applications but legacy applications utilize POSIX. Campaign Storage is a Near-POSIX File System using cloud storage for data and many POSIX file systems for metadata. Campaign Storage scales namespace metadata to trillions of files and billions of files in a single directory and files from 1 byte to Petabytes. This solution is now available commercially. This talk describes Campaign Storage motivation, design, and performance.

Speakers
GG

Gary Grider

High Performance Computing Division Leader, Los Alamos National Laboratory
As High Performance Computing (HPC) Division at Los Alamos National Laboratory I am responsible for all aspects of HPC technologies and deployment at Los Alamos. I am also a DOE lead in Data Storage technology funding and management of national storage projects such as the projects... Read More →


Thursday March 23, 2017 10:25am - 11:15am EDT
Paul Revere C
  Distributed

10:25am EDT

Next Generation File Replication System In GlusterFS - Rafi Kavungal Chundattu Parambil, Red Hat
Achieving availability and correctness, without compromising performance,
is always a challenge in designing new storage replication technologies.
In this presentation, we introduce the next-generation file replication
system in GlusterFS. Server-side replication and a flexible consistency
model enable us to take advantage of faster networks and cover a wide range
of needs, from being a highly available,
highly consistent system, to a high-performance eventually consistent
system depending on your use case.

Speakers
avatar for Rafi Kavungal Chundattu Parambil

Rafi Kavungal Chundattu Parambil

Software Engineer, Red Hat
Rafi K C from India. I am a C developer presently working on Gluster FS, a distributed file system project at Red Hat. Remote Direct Memory Access(RDMA), snapshot creation, data tiering are some of the areas of Gluster I have actively worked on. Improving the performance of the distributed... Read More →


Thursday March 23, 2017 10:25am - 11:15am EDT
Thomas Paine AB
  Software Defined

11:25am EDT

Infiniband Network Block Device (IBNBD) - Danil Kipnis, Profitbricks GmbH
Infiniband network block device (IBNBD) uses RDMA to transfer block IO bypassing SCSI sublayer. The driver implements client-side server memory management and requires only two infiniband messages per read or write IO on the wire. We take advantage of the infiniband "immediate" field to inform a receiver where the transferred data can be found. The per-cpu infiniband connections on client side allow for an IO-response to be processed on the same cpu, the IO was submitted on. The block MQ shared tags feature is used to fairly share the resource among multiple mapped devices. From the user perspective only remote device path and ib port gid of the server are required on client side for mapping, while no configuration on server is necessary. In the presentation we describe the simplistic design of the driver, its application in cloud infrastructure and compare it with existing solutions.

Speakers
avatar for Danil Kipnis

Danil Kipnis

Software Developer, Profitbricks GmbH
After graduating in Computer Science at TU-Berlin, Danil Kipnis worked as a research assistant at the Telecommunication Networks Group coauthoring several journal and conference papers focused on MAC protocols for wireless networks. In 2013 he filled his current position of linux... Read More →


Thursday March 23, 2017 11:25am - 12:15pm EDT
William Dawes AB

11:25am EDT

A Journey in the Containerization of Software Defined Storage Systems - Luis Pabon, CoreOS
The container revolution is upon us. Containerized applications can now be easily packaged, deployed, and managed on a distributed system like Kubernetes. Like any operating system, Kubernetes provides process and resource management, but it does this across multiple compute nodes instead of a single node. Developers innovating new features need to be aware of how this new data center operating system affects their applications. Although some of these applications can be containerized with ease, the process of containerizing and deploying existing distributed storage systems has not been as simple. This presentation will take a look at the lessons learned during the containerization and deployment of storage systems like GlusterFS and Ceph onto Kubernetes. We will also discuss how containerization of a distributed storage system has affected its access to system resources.

Speakers
avatar for Luis Pabon

Luis Pabon

Principal Software Engineer, CoreOS
Luis Pabón is a software engineer at CoreOS. Prior to joining CoreOS in November of 2016, he worked at Red Hat Storage, NetApp Advanced Technology Group, and at EMC on various storage products. He also previously presented at Vault in 2016 and 2015.


Thursday March 23, 2017 11:25am - 12:15pm EDT
Paul Revere AB
  Containers

11:25am EDT

Linux SMR Support Update - Damien Le Moal, Western Digital
Speakers
avatar for Damien Le Moal

Damien Le Moal

Director, Western Digital
Damien Le Moal manages the System Software Group within Western Digital Research. Damien is a regular contributor to Linux kernel block I/O stack and file system support for the ZBC and ZAC standards (zoned block devices) and is a maintainer of the kernel dm-zoned device mapper and... Read More →


Thursday March 23, 2017 11:25am - 12:15pm EDT
Paul Revere C

11:25am EDT

Ceph Snapshots: Diving into Deep Waters - Gregory Farnum, Red Hat
Ceph includes snapshot technology in most of its projects: the base RADOS layer, RBD block devices, and CephFS filesystem. This talk will explore the implementation of each, with discussion about the consequences and causes of design choices, as well as alternatives which have been discussed and experimented with. If you’ve ever wondered
1) what RADOS "self-managed” snapshots are,
2) why we tell people to ignore "pool snapshots” every time they come up,
3) how CephFS can generate snapshots at arbitrary directories, and if it’s hard,
4) why hard links make *everything* terrible,
5) how cluster size expectations can result in very different system design choices,
then this is the talk for you!

Speakers
GF

Gregory Farnum

Principal Software Engineer, Ceph, Red Hat
Greg Farnum is a long-standing member of the core Ceph development group, having joined the project as the third full-time engineer after graduating from Harvey Mudd College in 2009. Now a Red Hat employee, Greg has done major work on all components of the Ceph ecosystem and currently... Read More →


Thursday March 23, 2017 11:25am - 12:15pm EDT
Thomas Paine AB

12:15pm EDT

Lunch (Attendees on Own)
Thursday March 23, 2017 12:15pm - 2:15pm EDT
TBA

2:15pm EDT

SCSI Testing and Libiscsi - Ronnie Sahlberg, Google
SCSI testing and Libiscsi (Ronnie Sahlberg) - SCSI is a ubiquitous storage protocol. It is found all around us, ranging from large disk arrays to tiny USB sticks as well as in software such as the Linux kernel target and various virtualization tools. In this presentation, Ronnie Sahlberg will present a SCSI protocol test tool that comes as part of libiscsi and talk about how it can help you improve your SCSI target.

Ronnie will also talk about the libiscsi project. An open source project that implements a high performance iscsi client that can be used directly from applications such as QEMU.

Speakers
RS

Ronnie Sahlberg

Google
Ronnie Sahlberg, Google, is a storage hacker active in both file and block storage. He is the maintainer for both libiscsi and libnfs as well as a frequent contributor to tgt, samba, wireshark and others.


Thursday March 23, 2017 2:15pm - 3:05pm EDT
William Dawes AB
  Block
  • Experience Level Any

2:15pm EDT

Overlayfs Snapshots - Amir Goldstein, Ctera Networks
Overlayfs snapshots project aims to bring file level snapshots capability to any local file system.
The project utilizes overlayfs copy-file-on-write quality to maintain snapshots of a directory sub-tree.
In this BoF, Amir will talk about the technical details of the implementation and will show a demo.

https://github.com/amir73il/overlayfs/wiki/Overlayfs-snapshots

Speakers
avatar for Amir Goldstein

Amir Goldstein

Ctera Networks
Amir Goldstein is currently a filesystem developer at CTERA Networks (http://www.ctera.com). Prior to CTERA, Amir lead technology groups at various start-up companies in the fields of security, storage and containers. Amir is the creator and maintainer of out-of-tree Next3 project... Read More →


Thursday March 23, 2017 2:15pm - 3:05pm EDT
Paul Revere AB
  Containers

2:15pm EDT

Road Map of the Features in NFS v4.1, v4.2 and Beyond - Steve Dickson, Red Hat & Chuck Lever, Oracle
Chuck Lever and I will be talking about the currently implemented features in the v4.1 and v4.2 NFS protocols and outline the currently pending extensions to v4.2 that are working their way through the IETF process. Finally, we will talk about the IETF process which will be used to introduce additional extensions into the NFSv4 protocol.

Speakers
SD

Steve Dickson

Consulting Software Engineer, Red Hat
Steve Dickson is a kernel Engineer at Red Hat. For years he as been the tech lead for the NFS team. He maintains a number of upstream packages and is responsible for the NFS offerings in both the RHEL and Fedora distros. In prior lives he worked on DEC's TruCluster product as well... Read More →
CL

Chuck Lever

Linux Kernel Architect, Oracle Corporation
Chuck Lever has been a contributor to the Linux NFS implementation for nearly 20 years, working on such features as IPv6 support and NFS/RDMA. He has also published several NFS-related RFCs.


Thursday March 23, 2017 2:15pm - 3:05pm EDT
Paul Revere C
  Distributed

2:15pm EDT

Apache BookKeeper: A High Performance and Low Latency Cloud Storage Service - Sijie Guo, Twitter
Apache BookKeeper is a high performance and low latency storage that is designed for high performance write ahead logging. It was originally designed by Yahoo! Research to provide highly available logging services for namenode in HDFS. Since then it has grown beyond the original scope and is being used by different companies like Salesforce, Twitter, and Yahoo, as a fundamental building block to build replicated services like stream store, pub/sub messaging, database transaction log, and many more.

In this presentation, JV Jujjuri (Salesforce) and Sijie (Twitter) will give an overview of Apache BookKeeper, from motivation, architecture, and performance to community. In addition, they will also show production use cases of Apache BookKeeper.

Speakers
SG

Sijie Guo

Twitter
Currently work for Twitter on DistributedLog/BooKeeper. Apache BookKeeper PMC Chair. Previously work for Yahoo! on push notification system.


Thursday March 23, 2017 2:15pm - 3:05pm EDT
Thomas Paine AB

2:15pm EDT

Sponsor Showcase
Thursday March 23, 2017 2:15pm - 3:25pm EDT
President's Ballroom A+B

3:05pm EDT

Coffee Break
Thursday March 23, 2017 3:05pm - 3:25pm EDT
Courtyard Reception

3:25pm EDT

3:25pm EDT

Container Interfaces for Storage: Are We there Yet - James Bottomley, IBM Research
Many talks about containers start with Orchestration systems like Docker or Kubernetes. However, this one will look at the storage impacts on the actual in-kernel container API. With the addition of the superblock namespace (essentially a user namespace for the kernel to filesystem boundary) much of the stage is now set for fixing one of the biggest underlying container problems: that of translating unprivileged container writes into real filesystem uid/gids. This talk will examine how this system works, why it is necessary and what pieces still need to be added for orchestration systems to make use of it (yes, we'll also cover fully unprivileged Docker ... but only briefly).

Speakers
avatar for James Bottomley

James Bottomley

Distinguished Engineer, IBM
James Bottomley is a Distinguished Engineer at IBM Research where he works on Cloud and Container technology. He is also Linux Kernel maintainer of the SCSI subsystem. He has been a Director on the Board of the Linux Foundation and Chair of its Technical Advisory Board. He went to... Read More →


Thursday March 23, 2017 3:25pm - 4:15pm EDT
Paul Revere AB
  Containers

3:25pm EDT

NFS @ Scale: Worst. Protocol. Evar. (Except for all the Others) - Kevin Vigor, Facebook
NFS, despite being a 20 year old and much-maligned protocol, is critical infrastructure for distributed POSIX filesystems at Facebook. In this presentation, Kevin will discuss lessons learned from administering NFS infrastructure at very large scale. Strategies and tools for improving user experience will be presented, including a novel FUSE-based NFS client.

Speakers
KV

Kevin Vigor

Software Engineer, Facebook
Kevin is currently employed at Facebook working on the Gluster distributed filesystem and surrounding ecosystem. He was previously employed at Fusion-IO working on high-performance (and sadly proprietary) linux block storage drivers. Significant experience presenting to engineering... Read More →


Thursday March 23, 2017 3:25pm - 4:15pm EDT
Paul Revere C
  Distributed

3:25pm EDT

Hyper Converged Cache for Distributed Storage - Yuan Zhou, Intel
Hyper-converged system is becoming increasingly popular in IT infrastructure market, it integrates compute, storage, network and virtualization technologies into a single system.Traditional distribute storage system suffers from performance issues due to the remote data access to storage node and lack important features like deduplication, compression.
In this session, we would like to present a hyper converged cache solution for distributed storage systems. This hyper-converted cache solution employs a plugged framework, configurable cache mode like read-only, write-back and write through, yet provides strong data reliability, crash consistency, rich data services like deduplication and compression. Initial fio with zipf distribution benchmark results showed the hyper-converged cache demonstrated up to 30x performance improvement and 97% latency reduction with one NVMe SSD drive

Speakers
YZ

Yuan Zhou

Yuan Zhou is a Senior Software Development Engineer in the Software and Service Group for Intel Corporation, working in the Data Center and Software Technology team primarily focused on Cloud Storage Software. He has been working in Databases, Virtualization and Cloud computing for... Read More →


Thursday March 23, 2017 3:25pm - 4:15pm EDT
Thomas Paine AB

4:25pm EDT

SCSI-MQ on Legacy Hardware - Hannes Reinecke, SUSE Labs
linux block multiqueue has been implemented to provide best performance for modern multi-queued IO subsystems like NVMe. The SCSI stack also has been converted to make use of it by providing a simple 1-queue implementation.

However, most existing HBAs like LSI/Broadcom mpt3sas or PMC aacraid already employ MSI-X interrupt steering to spread out the load during interrupt processing, But this is only for the reception side, not the submission side.
So it doesn't quite match with the premises of block-mq, which assumes queue pairs, one for the submission and one for the reception side.

In this presentation I will be giving a performance analysis of the various methods on how block-mq can be mapped onto these HBAs, with the aim of providing tuning hints for the best possible performance and establish the best overall default parameters.

Speakers
avatar for Hannes Reinecke

Hannes Reinecke

Teamlead Storage & Networking, SUSE Linux GmbH
Studied Physics with main focus image processing in Heidelberg from 1990 until 1997, followed by a PhD in Edinburgh 's Heriot-Watt University in 2000. Working at SUSE Labs with focus on storage and mainframe. Principal contact point for storage related issues on SLES and teamlead... Read More →


Thursday March 23, 2017 4:25pm - 5:15pm EDT
William Dawes AB
  Block

4:25pm EDT

Overlayfs And Containers - Vivek Goyal & Miklos Szeredi, Red Hat
Overlayfs is a relatively new filesystem which implements copy on write semantics at the file level. It is specially gaining popularity in the container world where container technologies like docker are using overlayfs for image and container storage.

This talk plans to give an overview of overlayfs design, what are some of the outstanding concerns and what is being done to address those upstream. It also provides some details of how docker makes use of overlayfs and how it provided some significant speed up w.r.t
devicemapper. In the end the talk dives into details of SELinux support recently added to overlayfs, what were some of the challenges and how these were handled.

Speakers
VG

Vivek Goyal

Vivek Goyal works as a principal software engineer at Red Hat in Westford, MA,USA. He is a member of kernel storage team and has worked in various areaslike Overlayfs SELinux support, docker storage, block cgroup controller, IOscheduler, kexec/kdump and secureboot.Vivek has presented... Read More →
avatar for Miklos Szeredi

Miklos Szeredi

Senior Software Engineer, Red Hat
Miklos Szeredi is a Linux kernel hacker working for Red Hat. He has been interested in virtual filesystems for a long time, starting several open source projects including Filesystem in Userspace (FUSE) and the Overlay Filesystem.Prior to joining Red Hat, he has worked at SUSE Labs... Read More →


Thursday March 23, 2017 4:25pm - 5:15pm EDT
Paul Revere AB

4:25pm EDT

Provisioning NFSv4 Storage Using NFS-Ganesha, Gluster, and Pacemaker HA - Kaleb S. Keithley, Red Hat Gluster Storage
A deep dive into setting up a highly available NFSv4 cluster using NFS-Ganesha, Gluster and Gluster's libgfapi, and Pacemaker. NFS-Ganesha is a NFS server in user space that supports NFSv3, NFSv4, NFSv4.1, and pNFS. Kaleb will cover some Gluster internals, what libgfapi is and how it works, the NFS-Ganesha NFSv4 server and its File System Abstraction Layer (FSAL) model, and FSAL_GLUSTER in particular. He will show an example of how Pacemaker and Corosync can be used to create a resilient active/active cluster of servers where zero downtime is a requirement. He will describe at a high level how locking and lock recovery work in a cluster.

Speakers
KS

Kaleb S. Keithley

Sr. Principal Software Engineer, Red Hat
Kaleb currently works on Red Hat Gluster Storage and Community GlusterFS. Prior to the Gluster acquisition Kaleb worked on HekaFS, a multi-tenant storage proof-of-concept based on GlusterFS. Prior to coming to Red Hat he worked for EMC's Centera and Atmos divisions on a portable... Read More →


Thursday March 23, 2017 4:25pm - 5:15pm EDT
Paul Revere C
  Distributed
  • Experience Level Any

4:25pm EDT

BlueStore, A New Storage Backend for Ceph, One Year In - Sage Weil, Red Hat
BlueStore is a new storage backend for Ceph OSDs that consumes block devices directly, bypassing the local XFS file system that is currently used today. It's design is motivated by everything we've learned about OSD workloads and interface requirements over the last decade, and everything that has worked well and not so well when storing objects as files in local files systems like XFS, btrfs, or ext4. BlueStore has been under development for a bit more than a year now, and has reached a state where it is becoming usable in production. This talk will cover the BlueStore design, how it has evolved over the last year, and what challenges remain before it can become the new default storage backend.

Speakers
avatar for Sage Weil

Sage Weil

Ceph Project Leader, Red Hat
Sage Weil helped developed Ceph as part of his graduate studies at the University of California, Santa Cruz. After graduating he continued to develop the system and build an open source community around Ceph with the support of DreamHost. In 2012, he co-founded Inktank to productize... Read More →


Thursday March 23, 2017 4:25pm - 5:15pm EDT
Thomas Paine AB