moosefs部署Requirements | MooseFS Documentationmaster缓存分布式文件系统的源数据节点添加moosefs GPGKEYcurl https://repository.moosefs.com/RPM-GPG-KEY-MooseFS /etc/pki/rpm-gpg/RPM-GPG-KEY-MooseFScurl http://repository.moosefs.com/MooseFS-4-el9.repo /etc/yum.repos.d/MooseFS.reposerver:/var/lib/mfs 存储目录[rootserver1 yum.repos.d]# yum install moosefs-cli moosefs-gui -y[rootserver1 yum.repos.d]# systemctl enable --now moosefs-gui.serviceyum install moosefs-master 社区版这些示例文件复制到它们的实际配置名称中cd /etc/mfscp mfsmaster.cfg.sample mfsmaster.cfgcp mfsexports.cfg.sample mfsexports.cfgsystemctl enable --now moosefs-pro-master.servicechunkserver:[rootserver3 etc]# yum install moosefs-chunkserver -y复制默认配置模板cd /etc/mfscp mfschunkserver.cfg.sample mfschunkserver.cfgcp mfshdd.cfg.sample mfshdd.cfg添加一块硬盘[rootserver2 ~]# mkfs.xfs /dev/nvme0n1 -f 格式化[rootserver2 ~]# blkid[rootserver2 ~]# mount /dev/nvme0n1 /mnt/ssd01[rootserver2 ~]# vim /etc/fstab 为了将/dev/nvme0n1的挂载配置为永久生效clicchown -R mfs:mfs /mnt/ssd01[rootserver2 ~]# systemctl start moosefs-chunkserver.serviceclient[rootserver4 ~]# yum install moosefs-client建立MooseFS共享 自动挂载[rootserver4 ~]# mkdir /mnt/mfs[rootserver4 ~]# mfsmount /mnt/mfs -H mfsmaster或者添加默认挂载点cd /etc/mfsvim mfsmount.cfgmfsmountpacemaker高可用:转到负载均衡高可用篇server1和server5配置iscsi存储配置iscsizi输出端资源server4:添加一块硬盘[rootserver4 ~]# yum install -y targetcli[rootserver4 ~]# systemctl enable --now targettargetcli/ /backstores/block create my_disk /dev/sdb/ /iscsi create iqn.2026-07.com.example:storage1/ /iscsi/iqn.2026-07.com.example:storage1/tpg1/luns create /backstores/block/my_disk/ /iscsi/iqn.2026-07.com.example:storage1/tpg1/acls create iqn.2026-07.com.example:client/ exit配置iscsi输入端资源配置[rootserver1 ~]# yum install -y iscsi-initiator-utils.x86_64[rootserver5 ~]# yum install -y iscsi-initiator-utils.x86_64编辑配置文件 /etc/iscsi/initiatorname.iscsi写入InitiatorNameiqn.2026-07.com.example:server1当 server1 连接 iSCSI 目标时它会用这个 IQN 向 target 报告自己的身份在 server4 上为两个 IQN 分别添加 ACLtargetcli /iscsi/iqn.2026-07.com.example:storage1/tpg1/acls create iqn.2026-07.com.example:server1 /iscsi/iqn.2026-07.com.example:storage1/tpg1/acls create iqn.2026-07.com.example:server5 exit发现并登录 iSCSI 目标[rootserver1 ~]# iscsiadm -m discovery -t st -p192.168.80.154输出地址iscsiadm -m node -l分区格式化仅做一次server1[rootserver1 ~]# fdisk /dev/sdb[rootserver1 ~]# mkfs.xfs /dev/sdb1[rootserver1 ~]# mount /dev/sdb1 /mnt/mfs_data[rootserver1 ~]# chown mfs.mfs /mnt/[rootserver1 ~]# cp -p * /mnt/mfs_data[rootserver1 ~]# umount /mnt/mfs_data/[rootserver1 ~]# mount /dev/sdb1 /var/lib/mfs用作集群的数据存储[rootserver1 mfs]# systemctl stop moosefs-masterserver5:yum install moosefs-cli moosefs-gui moosefs-master -yll /dev/sdb1systemctl start moosefs-master保证server1和server5关闭开机自启moosefs-master服务pcs resource create mfs-data ocf:heartbeat:Filesystem \ device/dev/disk/by-partuuid/311a7dfc-01 \ directory/mnt/mfs_data \ fstypexfs \ optionsnoatime \ op start timeout60s \ op stop timeout60s \ op monitor interval10s[rootserver1 ~]# pcs resource create mfsmaster systemd:moosefs-master op monitor interval30s timeout20ssystemd:moosefs-master指定资源代理类型。Pacemaker 不会直接启动进程而是调用 Linux 系统的systemctl命令来管理moosefs-master.service这个系统服务[rootserver1 ~]# pcs resource group add mfscluster vip mfs-data mfsmaster创建一个名为mfscluster的资源组并按顺序将vip、mfs-data、mfsmaster放入组中实现先启动VIP再挂载存储最后启动mfsmaster主服务pacemaker fence_sbd1、创建 SBD 共享 vmdk 磁盘以管理员身份打开 Windows cmd进入 VMware 安装目录2、给每台集群虚拟机添加 SBD 磁盘所有虚拟机必须关机不能开机状态操作使用现有磁盘选择刚刚建的sdc保存现有格式删除快照才能永久设置找到虚拟机配置文件并修改按照路径打开相应虚拟机编辑配置文件disk.locking FALSEdisk.EnableUUID TRUEscsi1.sharedBus virtualscsi1:0.mode independent‑persistent多个100m磁盘