拓扑图
注意事项
1.各节点时间保持基本一致2.心跳ip一定要配置一．共享存储配置
Target配置1.修改用户名
[root@localhost~]#vim/etc/sysconfig/networkHOSTNAME=target.a.com
[root@localhost~]#hostnametarget.a.com2.修改解析文件hosts
[root@target~]#vim/etc/hosts192.168.4.1node1.a.com192.168.4.2node2.a.com
192.168.4.3target.a.com3.配置yum源
[root@target~]#vim/etc/yum.repos.d/rhel-debuginfo.repo[rhel-server]
name=RedHatEnterpriseLinuxserverbaseurl=file:///mnt/cdrom/Server/
enabled=1gpgcheck=0
gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-redhat-release[rhel-cluster]
name=RedHatEnterpriseLinuxclusterbaseurl=file:///mnt/cdrom/Cluster/
enabled=1gpgcheck=0
gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-redhat-release[rhel-clusterstorage]
name=RedHatEnterpriseLinuxclusterstoragebaseurl=file:///mnt/cdrom/ClusterStorage/enabled=1gpgcheck=0
gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-redhat-release4.更新软件源
[root@target~]#mount/dev/cdrom/mnt/cdrom
mount:blockdevice/dev/cdromiswrite-protected,mountingread-only[root@target~]#yumupdate~
5.安装target软件
[root@target~]#yuminstallscsi-target-utils6.创建共享磁盘
[root@target~]#fdisk/dev/sdb
DevicecontainsneitheravalidDOSpartitiontable,norSun,SGIorOSFdisklabelBuildinganewDOSdisklabel.Changeswillremaininmemoryonly,untilyoudecidetowritethem.Afterthat,ofcourse,thepreviouscontentwon'tberecoverable.
Thenumberofcylindersforthisdiskissetto1305.Thereisnothingwrongwiththat,butthisislargerthan1024,andcouldincertainsetupscauseproblemswith:
1)softwarethatrunsatboottime(e.g.,oldversionsofLILO)
2)bootingandpartitioningsoftwarefromotherOSs
(e.g.,DOSFDISK,OS/2FDISK)
Warning:invalidflag0x0000ofpartitiontable4willbecorrectedbyw(rite)Command(mforhelp):nCommandaction
eextended
p
primarypartition(1-4)
p
Partitionnumber(1-4):1
Firstcylinder(1-1305,default1):Usingdefaultvalue1
Lastcylinderor+sizeor+sizeMor+sizeK(1-1305,default1305):+1GCommand(mforhelp):w
Thepartitiontablehasbeenaltered!Callingioctl()tore-readpartitiontable.Syncingdisks.[root@target~]#7.配置scsi主文件
[root@target~]#vim/etc/tgt/targets.conf #ListoffilestoexportasLUNsbacking-store/dev/sdb1 #Authentication: #ifno\"incominguser\"isspecified,itisnotused#incominguserbackupsecretpass12 #Accesscontrol: #defaultstoALLifno\"initiator-address\"isspecifiedinitiator-address192.168.4.0/24 8.开启scsi服务 [root@target~]#servicetgtdstartStartingSCSItargetdaemon:9.查看配置 [root@target~]#tgtadm--lldiscsi--opshow--modetargetTarget1:iqn.2013-01.com.a:target Systeminformation: Driver:iscsiState:ready I_Tnexusinformation:LUNinformation: LUN:0 Type:controller SCSIID:deadbeaf1:0SCSISN:beaf10Size:0MB Online:Yes Removablemedia:No Backingstore:NobackingstoreLUN:1 Type:disk SCSIID:deadbeaf1:1 SCSISN:beaf11Size:1012MBOnline:Yes Removablemedia:NoBackingstore:/dev/sdb1 Accountinformation:ACLinformation: 192.168.4.0/24 Node2.a.com配置 [OK ] 1.配置如上，与target配置相同 [root@localhost~]#vim/etc/sysconfig/networkHOSTNAME=node2.a.com [root@localhost~]#hostnamenode2.a.com[root@node2~]#scp192.168.4.3:/etc/hosts /etc/ [root@node2~]#scp192.168.4.3:/etc/yum.repos.d/rhel-debuginfo.repo/etc/yum.repos.d/[root@node2~]#mount/dev/cdrom/mnt/cdrom mount:blockdevice/dev/cdromiswrite-protected,mountingread-only[root@node2~]#yumupdate2.安装软件iscsi [root@node2~]#yuminstalliscsi-initiator-utils3.配置iscsi用户名 [root@node2~]#vim/etc/iscsi/initiatorname.iscsiInitiatorName=iqn.2013-01.com.a.node24.启动iscsi服务 [root@node2~]#serviceiscsistartiscsidisstopped Turningoffnetworkshutdown.StartingiSCSIdaemon:SettingupiSCSItargets:iscsiadm:Norecordsfound! [OK ] 5.发现设备 [root@node2~]#iscsiadm--modediscovery--typesendtargets--portal192.168.4.3 192.168.4.3:3260,1iqn.2013-01.com.a:target6.登录设备 [root@node2~]#iscsiadm--modenode--targetnameiqn.2013-01.com.a:target--portal192.168.4.3:3260--login Logginginto[iface:default,target:iqn.2013-01.com.a:target,portal:192.168.4.3,3260] Loginto[iface:default,target:iqn.2013-01.com.a:target,portal:192.168.4.3,3260]:successful[root@node2~]#7.查看磁盘情况 [root@node2~]#fdisk-l Disk/dev/sda:21.4GB,21474836480bytes255heads,63sectors/track,2610cylindersUnits=cylindersof16065*512=8225280bytes DeviceBoot/dev/sda1/dev/sda2/dev/sda3 * Start1392546 End3825452610 Blocks Id System[OK ][OK ] 305203+83Linux20137477+83Linux 522112+82Linuxswap/Solaris Disk/dev/sdb:1011MB,1011677184bytes32heads,61sectors/track,1012cylinders Units=cylindersof1952*512=999424bytes Disk/dev/sdbdoesn'tcontainavalidpartitiontable8.在target上查看登录情况 [root@target~]#tgtadm--lldiscsi--opshow--modetargetTarget1:iqn.2013-01.com.a:target Systeminformation: Driver:iscsi State:ready I_Tnexusinformation: I_Tnexus:1 Initiator:iqn.2013-01.com.a.node2Connection:0 IPAddress:192.168.4.2 Node1.a.com配置 1.[root@localhost~]#vim/etc/sysconfig/networkHOSTNAME=node1.a.com [root@localhost~]#hostnamenode1.a.com [root@node1~]#scp192.168.4.3:/etc/hosts/etc/ [root@node1~]#scp192.168.4.3:/etc/yum.repos.d/rhel-debuginfo.repo/etc/yum.repos.d/[root@node1~]#mount/dev/cdrom/mnt/cdrom mount:blockdevice/dev/cdromiswrite-protected,mountingread-only[root@node1~]#yumupdate 2.安装软件 [root@node1~]#yuminstalliscsi-initiator-utils3.配置iscsi用户名 [root@node1~]#vim/etc/iscsi/initiatorname.iscsiInitiatorName=iqn.2013-01.com.a.node14.启动服务 [root@node1~]#servicescsicstartscsic:unrecognizedservice [root@node1~]#serviceiscsistart iscsidisstopped Turningoffnetworkshutdown.StartingiSCSIdaemon:SettingupiSCSItargets:iscsiadm:Norecordsfound! [ 5.发现设备 [root@node1~]#iscsiadm--modediscovery--typesendtargets--portal192.168.4.3192.168.4.3:3260,1iqn.2013-01.com.a:target6.登录设备 [root@node1~]#iscsiadm--modenode--targetnameiqn.2013-01.com.a:target--portal192.168.4.3:3260--login Logginginto[iface:default,target:iqn.2013-01.com.a:target,portal:192.168.4.3,3260] OK ] [OK ][ OK] Loginto[iface:default,target:iqn.2013-01.com.a:target,portal:192.168.4.3,3260]:successful7.查看磁盘情况[root@node1~]#fdisk-l Disk/dev/sda:21.4GB,21474836480bytes255heads,63sectors/track,2610cylinders Units=cylindersof16065*512=8225280bytesDeviceBootStart EndBlocksIdSystem/dev/sda1*138305203+/dev/sda239254520137477+/dev/sda3 2546 2610 522112+ Disk/dev/sdb:1011MB,1011677184bytes32heads,61sectors/track,1012cylinders Units=cylindersof1952*512=999424bytes8.在target上查看登录情况 [root@target~]#tgtadm--lldiscsi--opshow--modetargetTarget1:iqn.2013-01.com.a:target Systeminformation: Driver:iscsiState:ready I_Tnexusinformation: I_Tnexus:1 Initiator:iqn.2013-01.com.a.node2Connection:0IPAddress:192.168.4.2 I_Tnexus:2 Initiator:iqn.2013-01.com.a.node1 Connection:0 IPAddress:192.168.4.1 二．Gfs磁盘配置 我们在node1.a.com上对其进行配置1.PV阶段 [root@node1~]#pvcreate/dev/sdb Physicalvolume\"/dev/sdb\"successfullycreated查看pv信息 [root@node1~]#pvcreate/dev/sdb Physicalvolume\"/dev/sdb\"successfullycreated[root@node1~]#pvscan /dev/cdrom:openfailed:Read-onlyfilesystem Attempttoclosedevice'/dev/cdrom'whichisnotopen.PV/dev/sdblvm2[964.81MB] 83Linux83Linux82 Linuxswap/Solaris Total:1[964.81MB]/inuse:0[02.VG阶段 ]/innoVG:1[964.81MB] [root@node1~]#vgcreatevg0/dev/sdb /dev/cdrom:openfailed:Read-onlyfilesystem Attempttoclosedevice'/dev/cdrom'whichisnotopen.Clusteredvolumegroup\"vg0\"successfullycreated查看VG信息 [root@node1~]#vgscan [root@node1~]#vgscan Readingallphysicalvolumes.Thismaytakeawhile.../dev/cdrom:openfailed:Read-onlyfilesystemAttempttoclosedevice'/dev/cdrom'whichisnotopen.Foundvolumegroup\"vg0\"usingmetadatatypelvm23.LV阶段 [root@node1~]#lvcreate-L512-nlv0vg0Logicalvolume\"lv0\"created查看LV信息 [root@node1~]#lvscanACTIVE'/dev/vg0/lv0'[512.00MB]inherit查看生成的虚拟磁盘 [root@node1~]#ll/dev/vg0total0 lrwxrwxrwx1rootroot19Jan2419:31lv0->/dev/mapper/vg0-lv04.格式lv0为gfs格式 [root@node1~]#gfs_mkfs-plock_dlm-tcluster1:lv0-j3/dev/vg0/lv0Thiswilldestroyanydataon/dev/vg0/lv0.Areyousureyouwanttoproceed?[y/n]yDevice:Blocksize: FilesystemSize:Journals: ResourceGroups:LockingProtocol:LockTable:Syncing...AllDone [root@node1~]# 5.测试 因为node1与node2使用的/dev/sbd是挂载到target上的/dev/sdb1，所以在node1上做的操作，也会相应的出现在node2上，我们可以查看一下在node2上生成的虚拟磁盘 [root@node2~]#ll/dev/vg0 /dev/vg0/lv040963272438 lock_dlmcluster1:lv0 total0 lrwxrwxrwx1rootroot19Jan2419:36lv0->/dev/mapper/vg0-lv0和node1上是一模一样的 我们也可以查看两个节点的同步性 在Node1上挂载/dev/vg0/lv0,并且创建一个文件，查看node2上的情况[root@node1~]#mount/dev/vg0/lv0/var/www/html[root@node1~]#touch/var/www/html/index.html[root@node1~]#ll/var/www/html total4 -rw-r--r--1rootroot0Jan2419:41index.html在Node2上查看[root@node2~]#mount/dev/vg0/lv0/var/www/html[root@node2~]#ll/var/www/htmltotal4 -rw-r--r--1rootroot0Jan2419:41index.html可以看到达到了同步性 并且在其中一个节点中打开文件index.html时，另一个节点再打开这个文件是不可写的，这是锁机制。 三．RHCS配置 （1）Node2.a.com配置1.软件安装 [root@node2~]#yuminstall-yricci2.启动ricci服务 [root@node2~]#servicericcistartStartingoddjobd: generatingSSLcertificates...Startingricci: [ done [OK ]OK ] 3.安装httpd [root@node2~]#mount/dev/cdrom/mnt/cdrom（2）Node1.a.com配置软件安装riccilucihttpd [root@node1~]#yuminstall-yricciluci[root@node1cluster]#yuminstall-yhttpd1.启动ricci服务 [root@node1~]#servicericcistartStartingoddjobd: generatingSSLcertificates...Startingricci: [ done [OK ]OK ] [root@node1~]# 2.初始化luci，并启动服务 [root@node1~]#luci_admininit InitializingtheluciserverCreatingthe'admin'userEnterpassword:Confirmpassword:Pleasewait... Theadminpasswordhasbeensuccessfullyset.GeneratingSSLcertificates... TheluciserverhasbeensuccessfullyinitializedYoumustrestarttheluciserverforchangestotakeeffect.Run\"servicelucirestart\"todoso[root@node1~]# [root@node1~]#servicelucistart Startingluci:GeneratinghttpsSSLcertificates... done [ Pointyourwebbrowsertohttps://node1.a.com:8084toaccessluci3.安装httpd服务 [root@node1cluster]#yuminstall-yhttpd5.进入图形配置界面 在地址栏中输入ip:https://node1.a.com:8084 OK ] 或直接使用地址https://192.168.4.1:8084 5.创建一个群集 创建过程需要一段时间的等待创建之后会出现这个界面 创建fencedaemonkey 6.配置资源ip 7.配置资源http服务 8.配置资源存储与挂载 9.创建故障转移域 10.为节点指明fence 因为现在没有fence设备，所以我们可以为每个节点添加一个手工fence 11.添加服务 12.开启服务 13.验证开启 可以在两个节点上查看cluster的情况Node1.a.com上 [root@node2~]#clustat ClusterStatusforcluster1@ThuJan2420:19:552013MemberStatus:QuorateMemberName----------node2.a.comrgmanagernode1.a.comServiceNameState ----------------service:http-serverstarted Node1.a.com上 [root@node1~]#clustat ClusterStatusforcluster1@ThuJan2420:20:352013MemberStatus:QuorateMemberName----------node2.a.comnode1.a.comrgmanagerServiceNameState ----------------service:http-serverstarted 可以看出node1.a.com是主节点 IDStatus ----------1Online,Local,2Online,rgmanager Owner(Last)-----------node1.a.com ID Status ----------1Online,rgmanager 2Online,Local, Owner(Last)-----------node1.a.com 14.在两个节点中各创建一个网页主页 [root@node1~]#echo\"node1\">/var/www/html/index.html[root@node2~]#echo\"node2\">/var/www/html/index.html15.测试网页 16.断掉主节点 [root@node1~]#servicericcistopShuttingdownricci:查看情况 [OK] 刷新一个页面，会看到这种情况，这是因为我们没有fence设备，不能进行节点间的切换，所以我们不能得到node2.a.com为主节点 但是我们可以把node1.a.com节点关闭掉，开启节点node2.a.com测试下[root@node1~]#servicericcistart//先开启ricci服务Startingricci:[OK]关闭cluster服务 开启节点2 开启成功 测试网页 查看状态 [root@node2~]#clustat ClusterStatusforcluster1@ThuJan2420:19:552013MemberStatus:QuorateMemberName----------node2.a.comrgmanager ID----------1Online,Local,Status node1.a.comServiceNameState ----------------service:http-serverstarted 2Online,rgmanager Owner(Last)-----------node1.a.com 遇到的问题 在一个节点上创建vg0之后，在另一个节点上查看vg0状态时出现以下错误[root@node2~]#vgscan Readingallphysicalvolumes.Thismaytakeawhile... /dev/cdrom:openfailed:Read-onlyfilesystem Attempttoclosedevice'/dev/cdrom'whichisnotopen.Skippingclusteredvolumegroupvg0//skipping[root@node2~]#解决方案 在此节点下修改/etc/lvm/lvm.conf文件，修改locking_type的值为0然后用 vgchange-cnvg0 最后再把locking_type的值改回1 因篇幅问题不能全部显示，请点此查看更多更全内容