Who's using GlusterFS

From GlusterDocumentation

Please Note:

  • You can directly edit this page as a registered user or simply mail us the details to <whoisusing(at)gluster.com>.
  • You can describe in your own free style. No template or mandatory fields for submission.
  • It would be great if you could add the version you are using currently.
  • It is useful to provide as much as information (company name, address, hardware / software environment).
  • In case of confidentiality, you can at least provide the country name and what you are using GlusterFS for.
 PLACE HOLDER FOR NEW TESTIMONIALS 
 Oct 2009 | Comunip (Brazil)
 We are using GlusterFS 2.0 on Gentoo Linux distribution to provide high-availability VoIP services to our
 customers. Every server in the farm must be "hot-swap", and for this reason we needed a scalable, reliable and 
 flexible filesystem. GlusterFS was a very natural choice, due to its simplicity. Our internal development servers
 (mainly for file sharing and to host version control managers) benefits from its availability characteristics too.
 Jul 2009 Placenet.fr (FAI - France).
 Placenet.fr (FAI - FRANCE) use now in production glusterfs 2.0 on GNU/debian Lenny packages backported
 available here: | http://wunderland.placenet.fr/download. We use glusterfs in a farm of DomUs
 (clients glusterfs) and 2 Domus are use for glusterfsd (replication) in the case of mass mail and mass
 virtual hosting servers (cluster). the architecture will be available soon on our wiki (schema, 
 configuration files...)
August 2009 Andre Facina, São Paulo, SP, Brazil
http://www.jvsinfo.com.br
GlusterFS 2.0.2 on Red Hat Enterprise 5.2, 5.3 and Debian Lenny
Created a cluster + distributed replicated using IBM storage
ds3000 and ds4000. The client needed a high availability of information,
while a high performance. The Glusterfs solve the problem quickly and cost
an ideal.
July 2009 Edoceo, Seattle, WA, USA
http://edoceo.com/
GlusterFS 2.0.1 and 2.0.2 on Gentoo and Ubuntu

We use Gluster in-house to create a four-brick distributed+replicated 
multi- terabyte storage solution for ourselves.

Implemented multiple Gluster 2.x solutions for clients who have need for
highly reliable and resilient storage solutions.  Gluster 2.x has proven
itself to be highly dependable.  We have even used this to replace expensive
proprietary solutions.  There was much rejoicing!
April 2009 2L Multimedia, Annecy, FR, France
http://www.easyflirt.com/
GlusterFS 2.0rc4

We are the editor of Easyflirt, providing dating services. 
Since today we are testing GlusterFS in production environment, for caching
storage. After some months of testing and validating this solution, we will
use this storage for our members's pictures.

The setup consist of two servers Dell PE2950 with SAS drives. Eight PHP
servers are AFR clients and all performance translators. 
March 2009 Origo, Zurich, ZH, Switzerland
http://origo.ethz.ch
GlusterFS 2.0rc4

We have several servers providing the Origo online hosting platform.
Currently we use GlusterFS for storing log files from web and ftp servers
so that we can easily generate statistics. At a later point we might make
use of GlusterFS for other data as well.

For the log files we have a setup where two servers are storage backends
using DHS. We also use quota on the storage to restrict how much space
can be used.
January 2009 Ed Wyse Beauty Supply, Seattle, WA, USA
http://www.edwyse.com/
GlusterFS 2.0rc1

We have a small data center who's resources are underutilized. I'm testing 
an implementation of GlusterFS to maximize availability and redundancy.

My current test implementation uses 6 hosts with 4 drives in each host. 
Each drive is a brick. 1 brick from each of 3 hosts are joined with
cluster/replicate and each of those replicants are joined with
cluster/distribute. This gives us triple redundancy and can lose any
2 drives, or any 2 hosts without service interruption. I'll next be 
implementing cluster/ha for redundant data paths.
January  2009
http://www.consim.com//
We are internet based services company. Dedicated research team at Consim
takes up interesting technological challenges and paves way ahead for Consim 
by solving the same.
Similar to any other growing Internet services company, one striking 
challenge is SCALABILITY. In the process of addressing this important aspect,
we are benchmarking GlusterFS for our requirements and purpose. Results of same
may be posted in the  relevant section of this site soon.
 December 2008
 http://watson.nci.nih.gov/
 We are a research group within the US government using high-throughput genomic technologies to study 
 cancer.  The technologies that we use, particularly second-generation sequencers, generate a good bit
 of data (as much as 4TB or a little more per week, peak).  We had been using a largish NAS for data storage
 but noticed that our compute cluster was disk-limited (we had about 60 processors accessing the NAS 
 simultaneously).  We are hoping that glusterfs (1.3.12) will provide a flexible, higher performance 
 storage solution for less money and less complexity than a SAN.  Our glusterfs storage is limited, but 
 we expect to grow it.  Currently, we have about 35 TB of storage and a compute cluster of 60 processors.
 December 2008 [http://www.uwe.ac.uk University of the West of England, Bristol, UK
 Using GluserFS (1.3.12) over Infiniband to provide staff, students, researchers and research centers with
 storage space. Over 8000 users with around 600 user's reading from the storage at peak and about 150 users
 writing to the storage at peak. Feel free to view our deisgn diagram at:
                            http://www.bit.uwe.ac.uk/~jprsimmo/UWEStorageDesignv4.pdf
 Regards,
 John Simmonds (john2.simmonds@uwe.ac.uk)
 Bristol Institute of Technology
 December 2008 [http://www.viawest.net Viawest, Denver, Colorado, USA
 Currently testing glusterfs as alternative to SAN storage. Utilizing a cluster of recycled
 hosting customer servers, Viawest hopes to save over $500,000 USD compared to commercial vendors like
 Isilon or Ibrix. Our current deployment features GlusterFS 1.3.12 running on ArchLinux. Projections 
 show that with a small purchase of aprx $15,000 USD on replacement SATA Drives 
 (spanning 30 physical 1U servers), the total possible volume 135TB RAW, and aprx 65TB Useable
 when implementing AFR. Research and Testing has shown that GlusterFS is more flexible than Lustre or AFS.
 October 2008 inovex GmbH in Pforzheim, Germany
 We're using GlusterFS (1.3.12) in internal infrastructure production environment since October 2007 as 
 central storage for our backups (bacula), reexport over NFS and Samba and also for our FTP Server.
 Used Distribution is Debian Etch on 4 Storage Nodes (each with 8TB storage (16 x 500 Gig SATAII with Raid 6 
 and XFS)) and many Clients with client-side AFR and UNIFY, about 16 TB storage space.
 October 2008 BRIMIDES GmbH, QShare.com, Coburg, Germany
 Using GlusterFS (1.3.12) in production with our one-click hosting service.  Using four storage
 nodes and 2 clients on 2.6.18 Debian Etch 64bits, about 30TB storage space
 October 2008 Saskaweb IT Solutions in Saskatoon, SK, Canada.
 Using GlusterFS (1.3.12) in production with our offsite backup service.  Using two storage
 nodes and client-side AFR on 2.6.24
 October 2008 Tigron BVBA in Zaventem, Belgium.
 Using glusterFS (1.3.9) in production with billions of image files on two storage servers (sync)
 and three clients. Production platform is based on debian stable on Dell P2950 with Dell M1000 
 hardware storage enclosures. (+7 Terrabyte each) You can view the results each day at 
 http://www.delcampe.net.
 October 2008 CoreSpeed Hosting, Alpharetta, GA.
 Running GlusterFS (1.3.10) combined with Xen to create a scalable Grid Compute and Grid Filesystem setup.
 October 2008 Research group in Cosmology, Belgium.
 Testing GlusterFS (1.3.12) under Centos+ 5.2 on a small HPC open-source cluster: 6+1 nodes 
 of bi-quadcore CPUs. Using "unify NUFA" scheduler in single process mode in between the 6 computing
 nodes; and "ALU scheduler" in single process on the interactive node. Works fine till now!
 October 2008 Integrated Media Technologies, North Hollywood, CA
 Evaluating GlusterFS (1.4.0qa36) as a rich-media shared storage solution.  Testing performance 
 and functionality in heterogeneous environments, using OpenSolaris+ZFS for bricks and FCP stations
as clients.
                                 -- (Jose Palencia)
 September 2008 Goldeneye Solutions, Montreal, CA
 Testing GlusterFS and FreeBSD 7 to mirror user content over a small web cluster (4 nodes).
 September 2008 Materna GmbH, Germany
 We are testing GlusterFS to be used as afr/unify backend for file storage. Currently it`s being 
 used for a large cd-rom/iso repository, serving installation media for virtualized environments.
 If this works reliable, we will think about using it for other file-serving purpose.
 September 2008 Computer Vision Lab BIWI ETH Zürich, Switzerland
 We are currently testing GlusterFS to be employed on a small cluster of about 40 Nodes, each having 2x1TB. 
 We are very much looking forward to the BDB support, because we work with millions of small files (*.jpg)
 September 2008 Envoy Media Group - Tarzana, CA
 We have been using Gluster in production for 6 months now. Gluster shares files served by webservers
 along with php sessions and a slurry of config files. Moved from RedHat's GFS and never looked back. We 
 currently operate 2 storage nodes and 8 clients and are in the process of moving over 
 to Amazon's EC2 platform.
 September 2008 BC, Canada
 Evaluating Gluster to provide redundant storage for production active-active webserver cluster.
 September 2008 National newspaper, United Kingdom
 We have evaluated a number of options, and selected Gluster to provide shared storage 
 across our production web cluster.
 September 2008 Jyxo Czech Republic
 We are testing Gluster to store our databases from our Internet search engine crawler and
 other services which we provide to users.
 September 2008 Brightbox Rails Hosting
 We use GlusterFS on our production web clusters (with 6 to 20 nodes currently), storing hundreds 
 of gigs of mirrored assets for very busy web applications.  We're currently investigating its use
 as storage for our Xen and KVM virtual server disk images.
 September 2008 Mangahelpers
 We use GlusterFS on 4 nodes as HA storage for home directories, all website files,
 gentoo portage, bzr repository, and backups.  Currently in production.
 September 2008 Musmo Media
 GlusterFS will be the backbone of our cloud system; currently in testing. Very promising!
September 2008 ncibi.org
I installed gluster on a 180-core cluster that ncibi.org owns. In my opinion, glusterfs has a very
 neat design. For example, I ever used Lustre. Whenever I tried to create thousands of files, it just
failed. The cause is Luster is a modified ext3 file system, which is can only hold a limited number
of files under a folder. However Glusterfs doesn't need deal with that, because it uses fuse and
totally rely on back-end file system. Administrator can pick his favorite filesystem and use all
kinds of existing backup tools.
In summary, it is the best open source cluster file system, but I don't know if pNFS will give
glusterfs a hard time. Maybe I would not be loyal to glusterfs forever, :), but I do wish glusterfs
the best
                                 -- (Manhong Dai, University of Michigan)
September 2008 
I researched about a dozen file systems to run on Amazon's EC2 and couldn't find one I liked,
so I wrote my own, until I found glusterFS.  glusterFS was faster, more scalable, more flexible
and easy to build and install.  I had it running on EC2 in an hour.  Then I started chatting with
the Zresearch guys about adding support for Amazon S3 storage as a back end to glusterFS -- they
were incredibly responsive and helpful -- I wish all open source projects worked like this.
                                 -- (Mike Schroeder, DonorWare LLC, mike-gluster(at)donorware.com)
September 2008 
"I've spent a long time looking for a stable, reliable, simple and effective solution to aggregate
storage on various servers and one day I finally came across great guys at Zresearch... great work
guys, I wish you all the success you definitely deserve." 
                                 -- (Julien Perez, France / Slovakia, julien.perez(at)epsylonia.net)
 July 2008 sentimente.ro
 We use GlusterFS for our picture storage.  Currently in production.
 July 2008 FreeForm Technologies
 We use GlusterFS (currently 1.4.pre) for our cPanel based shared hosting solution.  Currently in production.
 July 2008 coull.com
 We use GlusterFS for our media storage solution, currently only in testing.
 July 2008 Siteground.com 
 We use GlusterFS for our Shared Hosting solutions, currently only in testing.
 July 2008 Archimedia s.n.c. - Brescia, Italy.
GlusterFS 1.3.9. Two nodes using server-side AFR for HA, exporting with multiple gluster instances mailboxes and web roots for users (~4TiB). Each node has 2x4cores, 8GiB of RAM and RAID5, networked to two Gigabit switches.
 June 2008 VA-Technologies - Luton, UK.
We are using GlusterFS 1.3.9 as a HA Storage Cluster.
8 Nodes, AFR'd, Unified. 4 Cores, 8Gb Ram & Raid. Nodes are bonded and trunked to Gigabit switch.
June 2008 Intergam Technologies Corporation - Chicago, IL, USA. 
We're using 4 nodes setup (unify over afr) for our internal storage.
May 2008 thebizmo / UGC Ltd. - London, UK - Media sales and distribution system.
Currently using a four node mirrored GlusterFS setup for content storage.
April 2008 OlyPen, Inc. - Sequim, WA, USA - Internet Service Provider and web hosting. 
 Currently testing AFR for web data replication, possibly moving to use with mail cluster.
Apr 2008 DKPSystem.com - Wisconsin, USA - Guild and clan hosting site. 
GlusterFS with AFR stores and serves user uploaded media. Three nodes currently.
Feb 2008 Curse, Inc. - MMORPG addon and social site.  We're using glusterfs 
AFR functionality to house all user generated content, enabling horizontal scaling of both download 
and upload servers. Currently two nodes, adding more in the coming months.
Feb 2008 Echo Systems, Inc. - Provider of metered online storage. We offer 
custom glusterfs server volumes from our highly available storage pool, starting at only USD 15¢/GB. 
Customer picks version.
Feb 2008 Universitat Internacional de Catalunya - Barcelona - Spain
[Universitat Internacional de Catalunya - http://www.uic.es/]
After nearly one year looking at this project, we're thinking to use it soon in our 
development environment. 
Just 4 bricks in an AFR configuration for a HD HA in a web farm cluster.
Oct 2007 GATC Biotech - Germany
Using GlusterFS as SAN Storage, based on 4 server bricks, serving a unified diskspace to
30 xen cluster clients. NFS was not stable enough for that high I/O we got. GLFS managed it
perfectly. 
Oct 2007 Indian Institute of Technology Kanpur - Chemistry department - India
24 brick GlusterFS storage on Infiniband for High Performance Computation Cluster
Oct 2007 Babcock & Brown - USA
AFR based GlusterFS storage on Infiniband High Performance Computation Cluster
Oct 2007 NAKAYAMA TECHNOLOGY CORPORATION 
www.nkym.com.ph - International Website
www.nkym.co.jp - Japan Website
We are going to use glusterfs for our file and web servers
26 Oct 2007 National Center for Supercomputing Applications (NCSA) - www.ncsa.uiuc.edu - United States
Testing GlusterFS as a possibility for shared mail spools.
26 Sep 2007 Nokta Internet Technologies - www.blogcu.com - Turkey
we using glusterfs tla , 2 file server and 15 high availability web farm (client) on linux.
We started to use glusterfs for failover and high speed, also it solved our io lock problem on 
our servers.  
24 Sep 2007 Cloud Nine - Testing a 5 brick GlusterFS instance as an alternative to using MogileFS
or our current NFS web. Virginia, USA.
23 Aug 2007 Unleash Computers Ltd - Nationwide, New Zealand
9 Aug 2007 Sunbeam Television (Video Archive Storage)
9 Jul 2007 mattsoftware.com, Australia
Currently experimenting with clusters for use in High Availability web farms.
We intend to use GlusterFS as the distributed filesystem to provide redundancy and failover.
9 Jul 2007 Stanford Exploration Project, Stanford University, California
GlusterFS storage on Infiniband for High Performance Computation Cluster
9 Jul 2007 WorldxChange Communications - Auckland, New Zealand.
8 Jul 2007 University of California, Merced - School of Engineering (Computer Vision, more specifically).
Jun 2007 Institute of Plasma Research - India
GlusterFS on Infiniband for High Performance Computation Cluster
1 Sept 2007 ww.com (testing and experiments, after having evaluated 'hadoop' and finding it unsuitable) 
16 Sep 2007 Yuhu.biz - BGServers.com - Buglaria 
We use GlusterFS for our Shared Hosting solutions on 3 servers which are on different
physical locations. 
GlusterFS 1.3.7, Kernels 2.6.23.1
Jan 2007 PDVSA R&D, PDVSA, Venezuela
GlusterFS on Infiniband for tape data migration and HPC cluster computing

 

Copyright © 2009 Gluster, Inc. All Rights Reserved.