GFS安装全过程.docx
GFS安装全过程GFS安装全过程#*8#*#*#&#Followthefollowingsequenceofreading:DatasharingwithaGESstoragecluster:/magazine006ar05features/gfs/RedHatClusterSuiteandManagingaClusterdocsmanualscsgfspdfrcfcs-en-4.pdfRedHatGFS6.1AdnliniStrator'sGuide/docs/manuals/csgfs/pdf/r_en-6_l.pdfG7S6.1ONRHAS4U2instalIationdocumentUster-4-howto.htmlDoanexperiment.#4#*Herearethreeofmymachineconfigurations,andthreePCSsetupasmal1environmenttest.Hostname,harddiskspaceforGFSspace,hosts192.168.100.124nociel192.168.100.1262192.168.100.125serverFdiSkisdistributedtoGFSSharedpartitions,butnotformattedInstallRpV,whichisdependent,sothereisasequenceAllthebagsinourofficialtraywerefound,andiftherewerenoavailabletodownload:/mirror,centos,org/centos-4/4.3csgfsi386RljMSTheRPMpackage:RPMS-ivhcman-kernel-2.66.50.2RPMS-ivhmile-1.7-1,i686.RPMRPM-ivhdim-kernel2.6.9-46.16RPMS-ivhCCS-1.0.10-0.i686.RPMRPM-ivhcman-1.0.17-0.i68.rpmRPM-ivhdlm-l.3-1.i686.RPMRPMivhgulm-1.0.1O-O.i686.RPMRPMS-ivh-plugins-1.0.12-0.i386.RPMRPMS-ivhperl-net-telnet-3.03-3.noarch.RPMRPM-ivhfence-1.32.45-1.i686.RPMRPMS-ivhipvsadm-1.24-6.i386.RPMRPM-ivhpiranha0.8.3T.i386.RPMRPM-ivhrgmanager-1.9.68-1.i386.RPMRPMS-ivhsystem-config-cluster-1.0.45-1.01.noarch.RPMRPMS-ivhGFS-kernel-2.6.9-72.2RPM-ivhGFS6.1.14-0.i386.RpMRPM-iv-ivhGNBD-kernel-2.6.9-10.20RPM-ivhGNBD-1.96-1.i686.RPMRPMS-ivhlvm2-cluster-2.02.21-7RPM-iv.iddev-2.0.0-4.i68.rpmClusterconfiguration/etc/cIuster/cIuster.Conf(key,ifyouhaveanXinterface,youcanconfigureitwithsystem-config-cluster)Bothnodesareconfigured?XM1.version=1.0?clusterconfig_version=1name=alpha_clusterfence_daemonpost_fail_delay=0post_join_delay=3/TheclusternodescIusternodename=nodelvotes=1Thefencemethodname=1devicename=f-ma11nodename=node1ipaddr=192.168.100.124/method/fence/clusternodecIusternodename=node2votes=1Thefence方法名=1设备名称=F-Man节点名=node2ipaddr=192.168.100.125/方法/栅栏/clusternode/clusternodescman/fencedevicesfencedevice剂=fence_manualname=F-Man/fencedevicesrmfailoverdomainsfaIloverdomainname=web_failover吩咐=1限制=Ofailoverdomainnodename=node1优先级=1/failoverdomainnodename=node2优先级=2/failoverdomain/faiIoverdoniains/rm/集群#»#»#相关服务：启动集群进程依次：冲洗液服务起先服务cman起先服务fenced起先服务clvmd起先服务gfs起先服务rgmanager起先停止服务的依次与启动的相反#*#*#*#在节点上加载dim模块rootnodel/#modprobelock_dlm启动ccsd服务roolodel/#冲洗液节点上启动集群管理器（Cmarorootnodel/#sbincman-tool加入冲洗液测试服务rootnodel/#CcsJest连接查看节点状态#猫proccluster/nodesrootnodel/力口入栅栏域：rootnodel/4/sbin/fence_tool力口入查看集群状态#猫,小0，/，】UStCr/statusrootnodel/在服务器端导出设备启动gnbd_serv进程rootserver/#/sbin/gnbd_serr-v-n导出设备rootserver/#gnbd_export-v-egfs-d/dev/hda4-c查看出口状态信息rootserver/#gnbd_export-v-1节点导入设备CrootQnodelttmodprobegnbd导入设备root©node1/gnbd_import-v-i-gnbd-server（此处的gnbd-server即是在/etc/主机里面服务器）查看导入状态信息（每个节点都要做）root0nodel/gnbd_import-v-1建立gfs文件系统并且挂载ErootSnodel/#modprobegfs在gnbd-server上建立gfs文件系统rootserver/#gfs_mkfs-talpha_cluster:gfs-j2/dev/gnbd/gfs（'2'表示有几个节点就写几,假如写错，得全部umount,来然后再重新起先modp）在每个节点挂载文件系统ErootQnodel/#mount-tgfs/dev/gnbd/gfs/gfstest（要在nodeO1挂载文件完全完成后，再在02年节点上进行山，否则会报错。得全部UmOUnt,来然后再重新起先modp）OK.done<.root©nodel/#df-hFilesystemSizeUsedvai1Use%Mountedon/devhda372G6.6G62G10%/devhdal99M8.8M85MbootNoneofthatis110M,0,HOMperhour,andthat,sthesamething/dev/GNBD/GFS/GFS/GFS/GFS/GFS/GFSrootnode2/#df-hFilesystemSizeUsedvai1Use%Mountedon/devhda369G6.6G,59G11%/devhdal99M8.8M85MbootNoneofthisis248M,0,248MdevSHM/dev/GNBD/GFS/GFS/GFS/GFS/GFS/GFS#»#»#Someonedidthefollowing,andIdidn,tencounteritduringtheexperiment.Thisseemstohavesomethingtodowithsomeofthethingsthatthebossgavelasttimeto1.VM.Activatethe1.VMvolumeRootone#vgchang-alyRoottow#vgchang-aly
