![\"GFS](\"http:\/\/rmohan.com\/wp-content\/uploads\/2015\/02\/GFS-001.jpg\")
<\/a><\/p>\n1. Add a cluster<\/p>\n
Log into the management interface, click Manage Clusters -> Create fill in the following:<\/p>\n
Cluster Name: gfs<\/p>\n
Cluster Name: gfs<\/p>\n NodeName \tPassword RicciHostname \t\t Ricci Port Select the following options, and then submit the Note: This step will create a cluster configuration file \/etc\/cluster\/cluster.conf<\/p>\n Fence Devices<\/p>\n Description: (1) log into the management interface, click Cluster-> Fence Devices -> 3. The device node binding Fence<\/p>\n Adding a node a fence<\/p>\n 1) Click cluster1.rmohan.com node, <\/p>\n Add Fence Method, here fill node01_fence; Description: VMNAME: virtual machine name, VM UUID: virtual machine.vmx file ” # \/usr\/sbin\/fence_vmware_soap -a 192.168.1.31 -z -l root -p test123 -n kvm_node2 -o list Adding a node two fence<\/p>\n 1) Click cluster2.rmohan.com node, Add Fence Method, here fill node02_fence; # Manual test fence function example: # Manual test fence function example:<\/p>\n # \/usr\/sbin\/fence_vmware_soap -a 192.168.1.31 -z -l root -p test123 -n kvm_node2 -o reboot Options: 4. Add Failover Domains Configuration<\/p>\n Name “gfs_failover” 5. Configure GFS Service<\/p>\n (1) GFS Service Configuration<\/p>\n In cluster1.rmohan.com, cluster2.rmohan.comstart CLVM services were integrated cluster lock<\/p>\n lvmconf –enable-cluster service clvmd start [root@cluster1 ~]# service clvmd start [root@cluster2 ~]# service clvmd start cluster1.rmohan.com<\/p>\n pvcreate \/dev\/sdb1<\/p>\n [root@cluster1 ~]# pvcreate \/dev\/sdb1 [root@cluster1 ~]# vgcreate gfsvg \/dev\/sdb1 # pvcreate \/dev\/sdc1 # pvs # vgcreate gfsvg \/dev\/sdb1 Failed to activate new LV. Note: just reboot the cluster2<\/p>\n [root@cluster1 ~]# lvcreate -l +100%FREE -n data gfsvg cluster2 (3) GFS file system format<\/p>\n cluster1 node:<\/p>\n [root@cluster1 ~]# mkfs.gfs2 -p lock_dlm -t gfs:gfs2 -j 2 \/dev\/gfsvg\/data Are you sure you want to proceed? [y\/n] y<\/p>\n Device: \/dev\/gfsvg\/data Description: 6. mount the GFS file system<\/p>\n cluster1, create a mount point on cluster2 GFS<\/p>\n [root@cluster1 ~]# mkdir \/vmdata [root@cluster2 ~]# mkdir \/vmdata<\/a><\/p>\n<\/a><\/p>\n
\ncluster1.rmohan.com test123 cluster1.rmohan.com 11111
\ncluster2.rmohan.com test123 cluster2.rmohan.com \t 11111<\/p>\n
\nUse locally installed packages.<\/p>\n
\nRHCS To achieve a complete clustering capabilities, we must realize fence function.
\nDue to the non-physical server configuration and other conditions, especially the use of VMware ESXi5.X virtual fence fence to realize the function of the device.
\nIt is thanks to the fence device can be used, was able to complete the test RHCS function.<\/p>\n
\n(2) Select the “Add selection VMware Fencing (SOAP Interface)
\n(3) Name “ESXi_fence”
\n(4) IP Address or Hostname “192.168.1.31” (ESXi Address)
\n(5) Login “root”
\n(6) Password “test123”<\/p>\n<\/a><\/p>\n<\/a><\/p>\n
\n2) add a fence instance, select “ESXi_fence” VMware Fencing (SOAP Interface)
\n3) VM NAME “kvm_cluster1”
\n4) VM UUID “564d6fbf-05fb-1dd1-fb66-7ea3c85dcfdf “check ssl<\/p>\n
\nuuid.location, “value, using the following format string.<\/p>\n
\nkvm_cluster2,564d4c42-e7fd-db62-3878-57f77df2475e
\nkvm_cluster1,564d6fbf-05fb-1dd1-fb66-7ea3c85dcfdf<\/p>\n
\n2) add a fence instance, select “ESXi_fence” VMware Fencing (SOAP Interface)
\n3) VM NAME “kvm_cluster2”
\n4) VM UUID “564d4c42-e7fd-db62-3878-57f77df2475e “check ssl<\/p>\n
\n<\/a><\/p>\n<\/a><\/p>\n
\nStatus: ON<\/p>\n
\n-o: List, status, and other parameters reboot<\/p>\n
\nPrioritized
\nRestricted
\ncluster1.rmohan.com 1
\ncluster2.rmohan.com 1<\/p>\n
\nchkconfig clvmd on<\/p>\n
\nActivating VG(s): No volume groups found [ OK ]<\/p>\n
\nActivating VG(s): 2 logical volume(s) in volume group “vg_cluster1” now active
\n [ OK ]
\n[root@cluster1 ~]#<\/p>\n
\nActivating VG(s): 2 logical volume(s) in volume group “vg_cluster2” now active
\n [ OK ]
\n[root@cluster2 ~]#<\/p>\n
\n Physical volume “\/dev\/sdb1” successfully created<\/p>\n
\n Clustered volume group “gfsvg” successfully created<\/p>\n
\n Physical volume “\/dev\/sdc1” successfully created<\/p>\n
\n PV VG Fmt Attr PSize PFree
\n \/dev\/sda2 vg_node01 lvm2 a– 39.51g 0
\n \/dev\/sdc1 lvm2 a– 156.25g 156.25g<\/p>\n
\n Clustered volume group “gfsvg” successfully created<\/p>\n
\n[root@cluster1 ~]# lvcreate -l +100%FREE -n data gfsvg
\n Error locking on node cluster2.rmohan.com: Volume group for uuid not found: EjHAOyOMJtk7pJ1gcUeOrbXjgFKMl05y2a3Mdh27oxKVpVBXjLYxHeU6088U9Ptc
\n Failed to activate new LV.<\/p>\n
\n clvmd not running on node cluster2.rmohan.com
\n Unable to drop cached metadata for VG gfsvg.
\n clvmd not running on node cluster2.rmohan.com<\/p>\n
\n# \/etc\/init.d\/clvmd start<\/p>\n
\nThis will destroy any data on \/dev\/gfsvg\/data.
\nIt appears to contain: symbolic link to `..\/dm-2′<\/p>\n
\nBlocksize: 4096
\nDevice Size 19.06 GB (4996096 blocks)
\nFilesystem Size: 19.06 GB (4996093 blocks)
\nJournals: 2
\nResource Groups: 77
\nLocking Protocol: “lock_dlm”
\nLock Table: “gfs:gfs2”
\nUUID: aaecbd43-cd34-fc15-61c8-29fb0c282279<\/p>\n
\ngfs: gfs2 gfs is the name of the cluster, gfs2 is to define the name, the equivalent of the label.
\n-j mount the file system specified number of hosts, you do not specify a default is the management node.
\nHere there are two nodes experiment<\/p>\n
\n[root@cluster1 ~]# mount.gfs2 \/dev\/gfsvg\/data \/vmdata
\n[root@cluster1 ~]#<\/p>\n
\n[root@cluster2 ~]# mount.gfs2 \/dev\/gfsvg\/data \/vmdata
\n[root@cluster2 ~]# df -TH
\nFilesystem Type Size Used Avail Use% Mounted on
\n\/dev\/mapper\/vg_cluster2-lv_root
\n ext4 19G 1.1G 17G 6% \/
\ntmpfs tmpfs 2.0G 33M 2.0G 2% \/dev\/shm
\n\/dev\/sda1 ext4 500M 52M 422M 11% \/boot
\n\/dev\/gfsvg\/data gfs2 21G 272M 21G 2% \/vmdata
\n[root@cluster2 ~]#<\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n