From GlusterDocumentation

Contents 1 Introduction 2 Example 2.1 node01 2.2 node02 2.3 node03 2.4 node04

Introduction

Earlier (before v1.3.8) GlusterFS needed two separate binary for both server part (glusterfsd) and client part (glusterfs). With v1.3.8, GlusterFS team came up with a single binary for both server part and client part. (now for compatibility sake, glusterfsd is a symlink to glusterfs).

Few important things why team developed this feature,

1. In NUFA kind of setup (where a machine is both client and server), the context switch over head due to two process was causing lot of performance bottleneck.
2. Ease of management with single process.
3. Saves lot of CPU cycle in forming protocol, memory allocation for that, and freeing the same.
4. Now, fuse is made as a translator.
5. This will reduce the number of packages in case of RPMs too (just single glusterfs binary RPM).
6. Running one command will start both server and client, hence simple entry in /etc/fstab, and mount/umount is enough to start and stop glusterfs.

Next sections shows the example of NUFA scheduler with Unify. But you can use this concept with any or all translators.

Example

This example considers that there are only 4 machines which are used as both client and servers. Also note that hostnames are used with current config, change it accordingly in your setup.

NOTE: If you are using AFR or stripe with single process spec file (ie, both client spec file and server spec file merged in one), note that the order of subvolumes in all nodes should remain constant.

node01

volume posix
 type storage/posix
 option directory /tmp/exports
end-volume

volume brick
 type performance/io-threads
 subvolumes posix
end-volume

volume brick-ns
 type storage/posix
 option directory /tmp/exports-namespace
end-volume

volume server
 type protocol/server
 option transport-type tcp/server
 option auth.ip.brick.allow *
 option auth.ip.brick-ns.allow *
 subvolumes brick brick-ns
end-volume

volume client-02
 type protocol/client
 option transport-type tcp/client
 option remote-host node02
 option remote-subvolume brick
end-volume

volume client-03
 type protocol/client
 option transport-type tcp/client
 option remote-host node03
 option remote-subvolume brick
end-volume

volume client-04
 type protocol/client
 option transport-type tcp/client
 option remote-host node04
 option remote-subvolume brick
end-volume

volume unify
 type cluster/unify
 option namespace brick-ns
 option scheduler nufa
 option nufa.local-volume-name brick
 option nufa.limits.min-free-disk 10%
 subvolumes brick client-02 client-03 client-04
end-volume

node02

volume posix
 type storage/posix
 option directory /tmp/exports
end-volume

volume brick
 type performance/io-threads
 subvolumes posix
end-volume

volume server
 type protocol/server
 option transport-type tcp/server
 option auth.ip.brick.allow *
 subvolumes brick
end-volume

volume client-01
 type protocol/client
 option transport-type tcp/client
 option remote-host node01
 option remote-subvolume brick
end-volume

volume client-03
 type protocol/client
 option transport-type tcp/client
 option remote-host node03
 option remote-subvolume brick
end-volume

volume client-04
 type protocol/client
 option transport-type tcp/client
 option remote-host node04
 option remote-subvolume brick
end-volume

volume client-ns
 type protocol/client
 option transport-type tcp/client
 option remote-host node01  # this is because we need a common namespace. you can choose to afr this.
 option remote-subvolume brick-ns
end-volume

volume unify
 type cluster/unify
 option namespace client-ns
 option scheduler nufa
 option nufa.local-volume-name brick
 option nufa.limits.min-free-disk 10%
 subvolumes brick client-01 client-03 client-04
end-volume

node03

volume posix
 type storage/posix
 option directory /tmp/exports
end-volume

volume brick
 type performance/io-threads
 subvolumes posix
end-volume

volume server
 type protocol/server
 option transport-type tcp/server
 option auth.ip.brick.allow *
 subvolumes brick
end-volume

volume client-01
 type protocol/client
 option transport-type tcp/client
 option remote-host node01
 option remote-subvolume brick
end-volume

volume client-02
 type protocol/client
 option transport-type tcp/client
 option remote-host node02
 option remote-subvolume brick
end-volume

volume client-04
 type protocol/client
 option transport-type tcp/client
 option remote-host node04
 option remote-subvolume brick
end-volume

volume client-ns
 type protocol/client
 option transport-type tcp/client
 option remote-host node01  # this is because we need a common namespace. you can choose to afr this.
 option remote-subvolume brick-ns
end-volume

volume unify
 type cluster/unify
 option namespace client-ns
 option scheduler nufa
 option nufa.local-volume-name brick
 option nufa.limits.min-free-disk 10%
 subvolumes brick client-01 client-02 client-04
end-volume

node04

volume posix
 type storage/posix
 option directory /tmp/exports
end-volume

volume brick
 type performance/io-threads
 subvolumes posix
end-volume

volume server
 type protocol/server
 option transport-type tcp/server
 option auth.ip.brick.allow *
 subvolumes brick
end-volume

volume client-01
 type protocol/client
 option transport-type tcp/client
 option remote-host node01
 option remote-subvolume brick
end-volume

volume client-02
 type protocol/client
 option transport-type tcp/client
 option remote-host node02
 option remote-subvolume brick
end-volume

volume client-03
 type protocol/client
 option transport-type tcp/client
 option remote-host node03
 option remote-subvolume brick
end-volume

volume client-ns
 type protocol/client
 option transport-type tcp/client
 option remote-host node01  # this is because we need a common namespace. you can choose to afr this.
 option remote-subvolume brick-ns
end-volume

volume unify
 type cluster/unify
 option namespace client-ns
 option scheduler nufa
 option nufa.local-volume-name brick
 option nufa.limits.min-free-disk 10%
 subvolumes brick client-01 client-02 client-03
end-volume