932
edits
Changes
m
Removing lab numbering to facilitate re-ordering
[[Category:OPS635]][[Category:OPS635-Labs]][[Category:peter.callaghan]]
=OPS635 Lab 4: Distributed FileSystems with GlusterFS=
==Overview==
In addition to traditional methods of remotely sharing files (e.g. NFS, samba), modern operating systems have the ability to access cloud-based distributed filesystems like GlusterFS. These can integrate the kinds of advantages you are used to seeing in RAID based fileystems (i.e. redundancy, self-healing, faster access) while removing the single-point-of-failure issue inherent in hosting a service on a single machine. While this lab does not explore the full functionality of this service (notably the use of arbiters, or combining the different types of GlusterFS volumes to gain the advantages of both), it will introduce you to be basics of creating a GlusterFS volume and accessing it from a client.
##Re-activate the brick, and observe that the client regains the ability to view all files.
==Investigation 2: Adding Bricks to a Distributed Filesystem in GlusterFS==
#On the third server, create a 1GiB brick and add it to the volume. ##Remember to stop the volume before you add the brick to it. Then restart it.#Use ls to observe even after it is added, no files are immediately stored on it.
##On the client create more files. Note that this may require re-mounting the filesystem in order to access the extended space.
==Investigation 3: Creating a Replicated Filesystem in GlusterFS==