GlusterFS is a popular, software-only distributed storage system and the lynchpin of the Gluster community. Every day, more users and developers come to appreciate the simplicity, ease of use, and flexibilty of scale-out storage, GlusterFS style. In this talk, attendees will learn about the project's history, what's new and what is coming just around the corner.
The Corosync/Pacemaker high availability stack is the leading HA environment on the Linux platform and will be the single supported HA stack in RHEL 7. This session outlines the current status of GlusterFS integration with Pacemaker, and ongoing improvements. Attendees will learn how to use GlusterFS as a distributed, scalable storage layer for Pacemaker clusters, as well as managing GlusterFS deployment with Pacemaker itself.
GlusterFS is a distributed file system that can scale to several PetaBytes. oVirt is a management platform for Kernel based Virtual Machine (KVM) and can be used to manage GlusterFS as well.
This presentation will discuss integration of KVM and GlusterFS through various mechanisms like:
Details on how both file and block based interfaces can be presented to host KVM images from GlusterFS will be provided.
The presentation will then talk about how oVirt can be used to provision GlusterFS volumes and how such volumes can then be used for hosting KVM images. Configuration details of both these features from oVirt would be presented.
The benefits emerging from integration of these projects would be highlighted as well.
Nowadays it's critical to keep information in a safe place whatever the contingency that may happen. However physical storage do fail and all its data can be lost. There exists multiple alternatives to mitigate this problem, like RAID1, RAID5 and some others. These were designed for local physical disks but the theory behind them can be exported to other scenarios.
GlusterFS has a built-in translator, called AFR or replicate, that implements the concept of a RAID1 but with a configurable number of replicas. Our disperse translator integrates the RAID5 concept into the GlusterFS stack, allowing a configurable level of reliability, i.e. support for a configurable number of failed nodes without loss of service. It's aimed to provide equivalent reliability to GlusterFS as AFR but using a fraction of the network bandwith and the storage capacity, using an optimized version of an IDA (Information Dispersal Algorithm).
This talk will focus on GlusterFS as a QEMU block backend which forms the basis of positioning GlusterFS as a storage backend for QEMU/KVM virtualization stack.
This talk will also cover Block Device translator for GlusterFS, which enables GlusterFS to work with block devices, the advantages, current limitations & status of this work and how this fits into the overall goal of making GlusterFS virtualization ready.
Virtualization and Storage administrators, GlusterFS developers as well as users of KVM virtualization will benefit from this talk.
Gluster experts Eco Willson and Niels de Vos will show off GlusterFS' ease of use and extensibility. Eco will kick things off with a quick start install and then proceed to a demonstration of our Unified File and Object (UFO) storage. UFO is a hybrid solution that allows you to access your existing data simultaneously via traditional filesystems and via the OpenStack SWIFT API's. Next, Niels de Vos will demonstrate how he integrated GlusterFS and Wireshark and how to add this integration to your security toolkit. Wireshark is a network protocol analyzer, letting you capture and interactively browse the traffic running on a computer network.
GlusterFS represents a dramatic departure from traditional backend storage solutions. In this talk, attendees will get a technical dive into GlusterFS from the SysAdmin perspective, including a study of implementation scenarios. We'll explore such topics as enterprise storage strategy, data access methods, the elegant simplicity of scaling both out and up, the strength of redundancy and fault tolerance, and ways to boost performance.
Gluster has a big users community and seems to be a good alternative to easily test any concept through its stackable translators approach. On the other hand, the current Distributed File Systems design has not really evolved and modern solutions still rely on an old metadata and I/O servers model. From our point of view, this model does not take into account the infrastructure’s physical topology, leading to expensive network communications that limit the performance and the scalability of the system. In this talk we'll discuss our impressions/advantages/problems of using Gluster when trying to develop a prototype that aims at reducing the traffic to the minimal needs and keep the applications performance over multiple sites.
