GlusterFS sur Rocky Linux 9

Configuration de GlusterFS pour le stockage distribué.

Composant	Version
Rocky Linux	9.x
GlusterFS	10+

Durée estimée : 40 minutes

---

Types de volumes

Type	Description	Min. serveurs
Distributed	Données réparties	2+
Replicated	Données dupliquées	2+
Dispersed	Erasure coding	3+
Distributed-Replicated	Combo	4+

---

Architecture

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│  Node 1     │     │  Node 2     │     │  Node 3     │
│  gluster01  │◄───►│  gluster02  │◄───►│  gluster03  │
│             │     │             │     │             │
│  /data/brick│     │  /data/brick│     │  /data/brick│
└─────────────┘     └─────────────┘     └─────────────┘
       │                   │                   │
       └───────────────────┼───────────────────┘
                          │
                   ┌──────▼──────┐
                   │   Client    │
                   │ /mnt/gluster│
                   └─────────────┘

---

1. Prérequis

Sur tous les nœuds

# Hostname
hostnamectl set-hostname gluster01  # 02, 03 sur les autres

# /etc/hosts
cat >> /etc/hosts << 'EOF'
192.168.1.11 gluster01
192.168.1.12 gluster02
192.168.1.13 gluster03
EOF

# Désactiver SELinux temporairement (ou configurer)
setenforce 0
sed -i 's/SELINUX=enforcing/SELINUX=permissive/' /etc/selinux/config

---

2. Installation

# Activer le repo CentOS Stream pour GlusterFS
dnf install -y centos-release-gluster10

# Installer
dnf install -y glusterfs-server

# Version
glusterfs --version

---

3. Préparer le stockage

# Partitionner le disque dédié
fdisk /dev/sdb
# Créer partition Linux (type 83)

# Formater en XFS
mkfs.xfs -i size=512 /dev/sdb1

# Monter
mkdir -p /data/brick1
echo '/dev/sdb1 /data/brick1 xfs defaults 0 0' >> /etc/fstab
mount -a

# Créer le répertoire brick
mkdir -p /data/brick1/gv0

---

4. Démarrer GlusterFS

systemctl enable --now glusterd
systemctl status glusterd

---

5. Firewall

firewall-cmd --permanent --add-service=glusterfs
firewall-cmd --reload

# Ou ports spécifiques
firewall-cmd --permanent --add-port=24007-24008/tcp
firewall-cmd --permanent --add-port=49152-49251/tcp
firewall-cmd --reload

---

6. Créer le cluster (Trusted Pool)

Depuis gluster01

# Ajouter les peers
gluster peer probe gluster02
gluster peer probe gluster03

# Vérifier
gluster peer status

---

7. Créer les volumes

Volume répliqué (mirror)

gluster volume create gv0 replica 3 \
    gluster01:/data/brick1/gv0 \
    gluster02:/data/brick1/gv0 \
    gluster03:/data/brick1/gv0

# Démarrer
gluster volume start gv0

# Info
gluster volume info gv0

Volume distribué

gluster volume create gv-dist \
    gluster01:/data/brick1/gv-dist \
    gluster02:/data/brick1/gv-dist \
    gluster03:/data/brick1/gv-dist

gluster volume start gv-dist

Volume distribué-répliqué (4 nœuds)

gluster volume create gv-distrep replica 2 \
    gluster01:/data/brick1/gv-distrep \
    gluster02:/data/brick1/gv-distrep \
    gluster03:/data/brick1/gv-distrep \
    gluster04:/data/brick1/gv-distrep

---

8. Monter côté client

Installation client

dnf install -y glusterfs-fuse

Montage

mkdir -p /mnt/gluster

# Montage direct
mount -t glusterfs gluster01:/gv0 /mnt/gluster

# Ou avec backup
mount -t glusterfs gluster01,gluster02:/gv0 /mnt/gluster

fstab

echo 'gluster01:/gv0 /mnt/gluster glusterfs defaults,_netdev,backup-volfile-servers=gluster02:gluster03 0 0' >> /etc/fstab

NFS (alternative)

# Activer NFS sur le volume
gluster volume set gv0 nfs.disable off

# Monter via NFS
mount -t nfs gluster01:/gv0 /mnt/gluster

---

9. Options de volume

Performance

# Cache
gluster volume set gv0 performance.cache-size 256MB
gluster volume set gv0 performance.io-thread-count 32

# Readdir-ahead
gluster volume set gv0 performance.readdir-ahead on

Sécurité

# Restreindre l'accès
gluster volume set gv0 auth.allow 192.168.1.*

# Refuser
gluster volume set gv0 auth.reject 192.168.1.100

Quotas

gluster volume quota gv0 enable
gluster volume quota gv0 limit-usage / 100GB
gluster volume quota gv0 list

---

10. Gestion

Status

gluster peer status
gluster volume status gv0
gluster volume info gv0

Ajouter un brick

# Préparation sur le nouveau nœud
gluster peer probe gluster04

# Ajouter (pour replica, multiple de replica count)
gluster volume add-brick gv0 replica 4 gluster04:/data/brick1/gv0

Supprimer un brick

gluster volume remove-brick gv0 replica 2 gluster03:/data/brick1/gv0 start
gluster volume remove-brick gv0 replica 2 gluster03:/data/brick1/gv0 status
gluster volume remove-brick gv0 replica 2 gluster03:/data/brick1/gv0 commit

Heal (réparation)

# Lancer le heal
gluster volume heal gv0

# Status
gluster volume heal gv0 info
gluster volume heal gv0 info healed
gluster volume heal gv0 info heal-failed

---

11. Snapshots

# Créer
gluster snapshot create snap1 gv0

# Lister
gluster snapshot list gv0

# Restaurer
gluster snapshot restore snap1

# Supprimer
gluster snapshot delete snap1

---

12. Monitoring

# Status détaillé
gluster volume status gv0 detail

# Profiling
gluster volume profile gv0 start
gluster volume profile gv0 info
gluster volume profile gv0 stop

# Top operations
gluster volume top gv0 read-perf
gluster volume top gv0 write-perf

---

Commandes utiles

# Cluster
gluster peer status
gluster peer probe <node>
gluster peer detach <node>

# Volumes
gluster volume list
gluster volume info [vol]
gluster volume status [vol]
gluster volume start/stop <vol>

# Maintenance
gluster volume heal <vol>
gluster volume rebalance <vol> start

---

Dépannage

# Logs
tail -f /var/log/glusterfs/glusterd.log
tail -f /var/log/glusterfs/bricks/*

# Connectivité
gluster peer status

# Heal status
gluster volume heal gv0 info

# Split-brain
gluster volume heal gv0 info split-brain

Problème	Solution
Peer rejected	Vérifier /etc/hosts, firewall
Split-brain	gluster volume heal gv0 split-brain
Volume offline	gluster volume start gv0 force

---

Changelog

Date	Modification
2024-12	Création initiale