UKUUG
(the UK's Unix & Open Systems User Group)
FAB 2008 — Files and Backup Seminars
Tuesday 19th & Wednesday 20th February
Provisional Programme
(In no particular order, exact topics and speakers may change and more talks may be added.)
Tuesday 19th February
Clustering with the GFS2 shared-storage file system
Steve Whitehouse, Red Hat UK Ltd.
GFS2 is a cluster file system for enterprise deployments. It allows you to manage clusters of servers, including large clusters (e.g. over 100 nodes) as if it was one server, allowing your normal applications to run without changes, without redundant copies and with simplified back-up and disaster recovery. (Slides).
An Introduction to High Availability
Chris Procter, Open Minds High Availability Solutions
High Availability clustering allows continuous access to your critical services. This presentation will explain the basics of High Availability clustering and what it can do for your servers, we will also look at the potential drawbacks and what you can do to overcome them.
How companies achieve High Availability
Shobana Patel, Open Minds High Availability Solutions
Server downtime can cause major disruptions in organisations where staff require 24/7 access to IT resources. This talk outlines High Availability Solutions and discusses customer stories, their particular problems, and how they have protected their applications and data. (Slides).
Secure Network filesystems with OpenAFS
Simon Wilkinson, University of Edinburgh
Edinburgh University is in the process of deploying a next generation, secure, networked filesystem for general computing use. This talk will present some of our experiences to date in this deployment, discussing local requirements for a networked file system, an evaluation of contenders in this space, and the detailed performance and stability investigation of AFS vs NFSv4 that we performed, and why AFS was chosen. (Slides).
Storage at (almost) the Speed of Light
Dr. Charles Curran, CERN (European Organization for Nuclear Research)
CERN, the world's largest particle physics laboratory will shortly start up its newest accelerator, the Large Hadron Collider. This will produce some 15PB of data every year, at rates up to 1.5GB/s (130 TB/day). This poses some interesting storage challenges for CERN's mass storage system and its IT personnel. Can we collect data at this rate? Where should we store it? Can the collected data be read back for analysis each year? Can it be kept more or less for ever? And, can this be done at a reasonable cost? Dr. Curran will talk about meeting these challenges. (Slides).
Wednesday 20th February - Backups & Recovery
Reliable Backups without tapes
Ben Harris, University of Cambridge
Historically, the usual approach to backing up computer systems involved using data tapes and storage in a fire safe. More recently, hard disks have become cheap enough that storing backups on spinning disks has become financially viable. For the last two years the University has been backing up many of its smaller Unix systems onto a pair of redundant, geographically diverse servers full of disks. This talk discusses what we did, how we did it and how well it worked out.
Orchestrate your Data Center
Jo De Baer, Adam Spiers, Novell
How can we have a tidy and efficient data centre, where physical failures are unnoticeable to end users, if our data centres contain heterogeneous hardware, operating systems, storage and other resources? Jo De Baer will look at storage management that is grid-inspired, usage-oriented and based on open standards such as CIM, SMASH, and SMI-S.
Backup and Recovery with Bacula
Kern Sibbald, Bacula (Lead Developer and founder of the Bacula project)
Bacula is the leading open-source network Backup Solution for Unix, Linux and Windows, it allows the system administrator to manage backup, recovery, and verification of computer data across a network of computers of different kinds. Kern will explain how to implement a Bacula-based backup and recovery system, from a handful of servers to hundreds of computers. (Slides).
Backup integration with filesystem snapshot capability
Howard Thomson, Gem Brian
Filesystems such as btrfs, Hammer, Ceph, and LVM are including snapshot, replication and rollback facilities as an integral part of their implementation. Integrating the implementation of a back-up system with the raw storage mechanisms of the file-system can both improve the performance of the back-up process and improve the degree of system consistency on restore.
Lightning Talks:
YOU!
Lightning talks allow delegates to present a five minute talk on any area covered by the FAB 2008 seminar, e.g. files, storage, backup, recovery, availability, etc. Book your slot beforehand by emailing: fab@ukuug.org – or volunteer on the day!
- For UK Unix Announcements
- Enter your email address:
Fax: 01763 273 255
Web: Webmaster
Queries: Ask Here
PO BOX 37
Buntingford
Herts
SG9 9UQ
More information