Introduction into Scale-out Storage
with Gluster
Niels de Vos
Red Hat Storage Engineer
GlusterFS co-maintainer
[email protected]
November 28, 2015
Eindhoven
Agenda
What is Gluster?
Architecture
Quick start
How to get involved?
2
What is Gluster ?
Gluster is a distributed scale out filesystem that allows
rapid provisioning of additional storage based on your
storage consumption needs. It incorporates automatic
failover as a primary feature. All of this is accomplished
without a centralized metadata server.
3
Gluster in Keywords
Scalable, general-purpose storage platform
POSIX-y Distributed File System
Object storage (swift)
Flexible storage (libgfapi)
No Metadata Server
Heterogeneous Commodity Hardware
Flexible and Agile Scaling
Capacity Petabytes and beyond
Performance Thousands of Clients
4
Gluster Use-Cases
Content Delivery Networks
Media Streaming
Download Servers
Archival
Backup services
Long term media archives
Virtual Machine images
High Performance / Distributed Computing
Rendering Farms
Big Data
5
Scale-out and Scale-up
6
Data Access Overview
GlusterFS Native Client
Filesystem in Userspace (FUSE)
NFS
Built-in Service, NFS-Ganesha with libgfapi
SMB/CIFS
Samba server required (libgfapi based module)
Gluster For OpenStack (Swift-on-file)
libgfapi flexible abstracted storage
Integrated with QEMU, Bareos and others
7
Architecture
8
Distributed Volume
Files evenly spread across bricks
Similar to file-level RAID 0
Server/Disk failure could be catastrophic
9
Replicated Volume
Copies files to multiple bricks
Similar to file-level RAID 1
10
Distributed Replicated Volume
Distributes files across replicated bricks
11
Quick start
Assuming you have a disk at /dev/sdb:
#fdisk/dev/sdb
Format the partition:
#mkfstxfs/dev/sdb1
Mount the partition as a Gluster "brick":
#mkdirp/bricks/testvol
#mount/dev/sdb1/bricks/testvol
12
Quick start
Add an entry to /etc/fstab:
#tailn1/proc/mounts>>/etc/fstab
Install Gluster packages on both nodes:
#yuminstallycentosreleasegluster
#yuminstallyglusterfsserver
13
Quick start
Run the gluster peer probe command:
#glusterpeerprobe<iporhostnameofsecondhost>
Configure your Gluster volume:
#glustervolumecreatetestvolrep2\
node01:/bricks/testvol/data\
node02:/bricks/testvol/data
Test using the volume:
#mkdir/mnt/gluster
#mounttglusterfsnode01:/testvol
#cpr/var/log/mnt/gluster
14
Recently added features
Arbiter Volumes
Policy based split-brain resolution
Bit-Rot detection
Sharding Volumes
Tiering
High-Availability for NFS-Ganesha with Pacemaker
Trashcan to recover deleted files
Glusterfind API
...
15
Upcoming Features for next releases
New Style Replication
Kerberized Gluster network transport
SElinux contexts on Gluster mountpoints
Improved support for sparse files
Steps towards simultaneous Samba and NFSv4 access
Common high availability configuration (Pacemaker)
Leases/Delegations
RichACL
...
16
How to get involved ?
Homepage
http://gluster.org/
Community IRC Chat (on Freenode)
#gluster (for general topics)
#gluster-dev (for developers)
#gluster-meeting (meeting room)
Mailing Lists
http://www.gluster.org/mailman/listinfo/gluster-users
http://www.gluster.org/mailman/listinfo/gluster-devel
Documentation
http://gluster.readthedocs.org/
https://access.redhat.com/ - Red Hat Gluster Storage
17
Thanks!
18
