Centos7 build GlusterFS

Centos7 build GlusterFS

Lab requirements:

· Four machines are installed with GlusterFS to form a cluster

· The client stores docker registry in the file system

· The hard disk space of the four nodes is not integrated into one hard disk space. Each node must be stored in one copy to ensure data security.

Environment Planning

Server

Node1: 192.168.0.165 Host Name: glusterfs1

Node2: 192.168.0.157 Host Name: glusterfs2

Node3: 192.168.0.166 Host Name: glusterfs3

Node4: 192.168.0.150 Host Name: glusterfs4

Client

192.168.0.164 Host Name: master3

Prerequisites

· All Hosts close the firewall and SElinux

· Modify the hosts file to parse each other

192.168.0.165glusterfs1 192.168.0.157glusterfs2 192.168.0.166glusterfs3 192.168.0.150glusterfs4 192.168.0.164master3

Install

Server

1. Install the GlusrerFS package on the glusterfs {1-4} Node

#wget-P/etc/yum.repos.d #yuminstall-yglusterfsglusterfs-serverglusterfs-fuse #servicegluterdstart #chkconfiggluterdon

2. Configure the entire GlusterFS cluster on the glusterfs1 node and add each node to the cluster.

[root@glusterfs1~]#glusterpeerprobeglusterfs1 1peerprobe：success：onlocalhostnotneeded [root@glusterfs1~]#glusterpeerprobeglusterfs2 1peerprobe：success [root@glusterfs1~]#glusterpeerprobeglusterfs2 1peerprobe：success [root@glusterfs1~]#glusterpeerprobeglusterfs2 　1peerprobe：success

3. View node status

[root@glusterfs1~]#glusterpeerstatus

4. Create a data storage directory on glusterfs {1-4}

#mkdir-p/usr/local/share/models

5. Create a GlusterFS disk on glusterfs1

Note:

Replica 4 is one of the four nodes, each of which stores data once, that is, four copies of data and one copy of each node.

If replica 4 is not added, the disk space of the four nodes is integrated into a hard disk,

[root@glusterfs1~]#glustervolumecreatemodelsreplica4glusterfs1:/usr/local/share/modelsglusterfs2:/usr/local/share/modelsglusterfs3:/usr/local/share/modelsglusterfs4:/usr/local/share/modelsforce 1volumecreate：models：success：pleasestartthevolumetoaccessdata

6. Start

[root@glusterfs1~]#glustervolumestartmodels

Client

1. Deploy the GlusterFS client and mount the GlusterFS File System

[root@master3~]#wget-P/etc/yum.repos.d [root@master3~]#yuminstall-yglusterfsglusterfs-fuse [root@master3~]#mkdir-p/mnt/models [root@master3~]#mount-tglusterfs-oroglusterfs1:models/mnt/models/

2. view results

[root@master3~]#df-h FilesystemSizeUsedAvailUse%Mountedon /dev/vda3289G5.6G284G2%/ devtmpfs3.9G03.9G0%/dev tmpfs3.9G80K3.9G1%/dev/shm tmpfs3.9G169M3.7G5%/run tmpfs3.9G03.9G0%/sys/fs/cgroup /dev/vda11014M128M887M13%/boot glusterfs1:models189G3.5G186G2%/mnt/models

Other operation commands

Delete A GlusterFS Disk

# Glustervolumestopmodels stop first # Delete glustervolumedeletemodels

Detach a GlusterFS Disk

glusterpeerdetachglusterfs4

ACL Access Control

glustervolumesetmodelsauth.allow10.60.1.*,10.70.1.*

Add a GlusterFS Node

#glusterpeerprobesc2-log5 #glusterpeerprobesc2-log6 #glustervolumeadd-brickmodelssc2-log5:/data/glustersc2-log6:/data/gluster

Migrate GlusterFS data

#glustervolumeremove-brickmodelssc2-log1:/usr/local/share/modelssc2-log5:/usr/local/share/modelsstart #glustervolumeremove-brickmodelssc2-log1:/usr/local/share/modelssc2-log5:/usr/local/share/modelsstatus #glustervolumeremove-brickmodelssc2-log1:/usr/local/share/modelssc2-log5:/usr/local/share/modelscommit

Fix GlusterFS data (when Node 1 is down)

#glustervolumereplace-brickmodelssc2-log1:/usr/local/share/modelssc2-log5:/usr/local/share/modelscommit-force #glustervolumehealmodelsfull