Eadata一体机使用的50个小技巧

TOC\o"1-5"\h\zHYPERLINK\l"bookmark0"\o"CurrentDocument"Exadata管理5...HYPERLINK\l"bookmark2"\o"CurrentDocument"Exadata性能优化5...HYPERLINK\l"bookmark4"\o"CurrentDocument"让表使用flashcache5可以使用如下公式计算Exadata特性对IO的优化5.HYPERLINK\l"bookmark6"\o"CurrentDocument"可以使用如下公式计算ExadataStorageIndex对DiskIO减少的共享5可以使用如下计算FlashCache的使用率5.收集cell级别的表缓存统计信息的方法6.HYPERLINK\l"bookmark8"\o"CurrentDocument"确认在使用writebackflashcache6.HYPERLINK\l"bookmark10"\o"CurrentDocument"确认所有的griddisk均为正常online状态7.HYPERLINK\l"bookmark12"\o"CurrentDocument"确认所有的flashdisk均为正常online状态7.HYPERLINK\l"bookmark14"\o"CurrentDocument"启用writebackflashcache的方法7.确认Exadata计算节点间的网络带宽9..HYPERLINK\l"bookmark16"\o"CurrentDocument"多个ORACLE_HOME是否RDS可用？9.HYPERLINK\l"bookmark18"\o"CurrentDocument"relinkORACLE_HOME的RDS1.0不同配置Exadata的推荐最大并行度1.0HYPERLINK\l"bookmark20"\o"CurrentDocument"ExadataEHCC支持1..0HYPERLINK\l"bookmark22"\o"CurrentDocument"Exadata压缩信息1..0针对写日志redo特别多的应用启用SmartFlashlogging特性11HYPERLINK\l"bookmark24"\o"CurrentDocument"ExadataDB管理1..1HYPERLINK\l"bookmark26"\o"CurrentDocument"Exadata存储空间计算1..1查看cell软件版本1..1了解cell的温度1..1cell存储节点的日志存放位置1.2HYPERLINK\l"bookmark28"\o"CurrentDocument"列出cell中的alerthistory1..2为cell创建一个告警阈值1..2HYPERLINK\l"bookmark30"\o"CurrentDocument"cell可用性监控1..2HYPERLINK\l"bookmark32"\o"CurrentDocument"如何禁用SmartScan?1..3HYPERLINK\l"bookmark34"\o"CurrentDocument"如何禁用storageindex?13HYPERLINK\l"bookmark36"\o"CurrentDocument"如何禁用flashcache?13cell相关的数据库视图有以下这些视图1.3HYPERLINK\l"bookmark38"\o"CurrentDocument"配置Inter-DatabaseIORM1.3HYPERLINK\l"bookmark40"\o"CurrentDocument"如何禁用布隆过滤BloomFliter1.4HYPERLINK\l"bookmark42"\o"CurrentDocument"Exadata数据备份1..4HYPERLINK\l"bookmark44"\o"CurrentDocument"backup备份速率1..4HYPERLINK\l"bookmark46"\o"CurrentDocument"recovery应用日志恢复速率1.4HYPERLINK\l"bookmark48"\o"CurrentDocument"standbydatabase搭建1..5HYPERLINK\l"bookmark50"\o"CurrentDocument"Exadata恢复1..5HYPERLINK\l"bookmark52"\o"CurrentDocument"cell救护1..5HYPERLINK\l"bookmark54"\o"CurrentDocument"Exadata部署1..5HYPERLINK\l"bookmark56"\o"CurrentDocument"onecommand下载1..5HYPERLINK\l"bookmark58"\o"CurrentDocument"Exadata安装前准备工作1..5环境检查1..6.验证网络连通性1..6exachk健康检查脚本1..8HYPERLINK\l"bookmark60"\o"CurrentDocument"Exadata文档信息1..9HYPERLINK\l"bookmark64"\o"CurrentDocument"Exadata硬件篇1..9.常规1..9.HYPERLINK\l"bookmark66"\o"CurrentDocument"默认密码，以下是Exadata中cell/dbnodeIB等的默认密码：19硬件常规巡检：2..0检测Exadata数据库机器上的硬件和固件版本是否匹配？20检测软件版本与平台是否匹配？2.0为cell启用告警2..1监控磁盘故障2..1HYPERLINK\l"bookmark68"\o"CurrentDocument"更换StorageCell硬盘2..1HYPERLINK\l"bookmark70"\o"CurrentDocument"观察DatabaseServer磁盘状态2.1HYPERLINK\l"bookmark72"\o"CurrentDocument"观察DatabaseServerRAID状态2.2HYPERLINK\l"bookmark74"\o"CurrentDocument"StorageCell加电启动2..2HYPERLINK\l"bookmark76"\o"CurrentDocument"检测memoryECC错误2..2若发现Exadata上存在磁盘损毁则：2.3HYPERLINK\l"bookmark78"\o"CurrentDocument"检测cellserverCachePolicy23HYPERLINK\l"bookmark80"\o"CurrentDocument"Exadata停机：2..5HYPERLINK\l"bookmark82"\o"CurrentDocument"Exadata启动2..6HYPERLINK\l"bookmark84"\o"CurrentDocument"Infiniband篇2..7HYPERLINK\l"bookmark86"\o"CurrentDocument"启停IBSwitch2..7查看IB网络拓扑状态2..8诊断IB链路没有错误2..8查看IB网络连线2..8查看IB健康状态2..8IB健康检查2..8IB故障处理2..9IB硬件监控2..9HYPERLINK\l"bookmark88"\o"CurrentDocument"Switch端口错误2..9HYPERLINK\l"bookmark90"\o"CurrentDocument"Link状态3..0HYPERLINK\l"bookmark92"\o"CurrentDocument"Subnetmanager3..0HYPERLINK\l"bookmark94"\o"CurrentDocument"CISCO交换机3..0.例行维护操作3..0.HYPERLINK\l"bookmark96"\o"CurrentDocument"KVM3..1.HYPERLINK\l"bookmark98"\o"CurrentDocument"PDU3..1.故障处理3..1.Exadata管理Exadata性能优化让表使用flashcacheALTERTABLEvobjectname>storage(CELL_FLASH_CACHEKEEP);可以使用如下公式计算Exadata特性对10的优化[1-{(cellphysicalIOinterconnectbytesreturnedbysmartscan)/(cellIOuncompressedbytes+cellphysicalIObytessavedbystorageindex)}]*100可以使用如下公式计算ExadataStorageIndex对DiskIO减少的共(cellphysicalIObytessavedbystorageindex/physicalreadtotalbytes)*100可以使用如下计算FlashCache的使用率(cellflashcachereadhit/physicalreadtotalIOrequests)*100收集cell级别的表缓存统计信息的方法SQL>SELECTdata_object_idFROMDBA_OBJECTSWHEREobject_name='EMP';OBJECT_ID57435CellCLI>LISTFLASHCACHECONTENT-WHEREobjectNumber=57435DETAILcachedSize:495438874dbID:70052hitCount:415483missCount:2059objectNumber:57435tableSpaceNumber:1确认在使用writebackflashcache#dcli-g~/cell_group-lrootcellcli-e“listcellattributesflashcachemode”Results:flashCacheMode:WriteBack->writebackflashcacheisenabledflashCacheMode:WriteThrough->writebackflashcacheisnotenabled确认所有的griddisk均为正常online状态#dcli-gcell_group-lrootcellcli-elistgriddiskattributesasmdeactivationoutcome,asmmodestatus确认所有的flashdisk均为正常online状态#dcli-gcell_group-lrootcellcli-elistflashcachedetail启用writebackflashcache的方法A.EnableWriteBackFlashCacheusingaROLLINGmethod(RDBMS&ASMinstaneeisup-enablingwrite-backflashcacheonecellatatime)Logontothefirstcellthatyouwishtoenablewrite-backFlashCacheDroptheflashcacheonthatcellcellcli-edropflashcacheCheckifASMwillbeOKifthegriddisksgoOFFLINE.Thefollowingcommandshouldreturn‘Yes'forthegriddisksbeinglisted:cellcli-elistgriddiskattributesname,asmmodestatus,asmdeactivationoutcomeInactivatethegriddiskonthecellcellcli-ealtergriddiskallinactiveShutdowncellsrv#cellcli-ealtercellshutdownscellsrvSetthecellflashcachemodetowriteback#cellcli-e“altercellflashCacheMode=writeback”Restartthecellsrv#cellcli-ealtercellstartupscellsrvReactivatethegriddisksonthecellcellcli—altergriddiskallactiveVerifyallgriddiskshavebeensuccessfullyputonlineusingthefollowingcommand:cellcli-elistgriddiskattributesname,asmmodestatusRecreatetheflashcachecellcli-ecreateflashcacheallCheckthestatusofthecelltoconfirmthatit'snowinWriteBackmode:cellcli-elistcelldetail|grepflashCacheModeRepeatthesesamestepsagainonthenextcell.However,beforetakinganotherstorageserveroffline,executethefollowingmakingsure‘asmdeactivationoutcome'displaysYES:#cellcli-elistgriddiskattributesname,asmmodestatus,asmdeactivationoutcomeB.EnableWriteBackFlashCacheusingaNON-ROLLINGmethod(RDBMS&ASMinstancesaredownwhileenablingwrite-backflashcache)Droptheflashcacheonthatcell#cellcli-edropflashcacheShutdowncellsrv#cellcli-ealtercellshutdownscellsrvSetthecellflashcachemodetowriteback#cellcli-e“altercellflashCacheMode=writeback”Restartthecellsrv#cellcli-ealtercellstartupscellsrvRecreatetheflashcache#cellcli-ecreateflashcacheall确认Exadata计算节点间的网络带宽可以采用ncnc-1.84-10.fc6.x86_64.rpm获得检测多个ORACLE_HOME是否RDS可用？dcli-g/opt/oracle.SupportTools/onecommand/dbs_group-loraclemd5sum${ORACLE_HOME}/lib/libskgxp11.sorelinkORACLE_HOME的RDSdcli-g/opt/oracle.SupportTools/onecommand/dbs_group-loracle“exportORACLE_HOME=$ORACLE_HOME;;cd'pwd';;make-fi*mkipc_rds”dcli-g/opt/oracle.SupportTools/onecommand/dbs_group-loracle“exportORACLE_HOME=$ORACLE_HOME;;cd'pwd';;make-fi*mkioracle”|egrep‘rm|mv.*oracle'不同配置Exadata的推荐最大并行度配置CPU个数推荐最大ParallelismFullRack64coreDOP=256HalfRack32coreDOP=128QuarterRack16coreDOP=64ExadataEHCC支持Exadata的EHCC支持宽表最大支持1000个字段的表，而不像11.1中的压缩仅支持最多255列的表Exadata压缩信息通过dbms_compression.get_compression_ratio可以获得表的压缩信息针对写日志redo特别多的应用建议启用SmartFlashlogging特性CREATEFLASHLOGALLCREATEFLASHLOGALLSIZE=1GCREATEFLASHLOGCELLDISK='fd1,fd2'CREATEFLASHLOGCELLDISK='fd1,fd2'SIZE=1GExadataDB管理Exadata存储空间计算FreeMB（最大可用空间）=GridDisk*12*NumofCells/RedundancyUsableMB（支持1个CELL故障的最大可用空间）=GridDisk*12*（NumofCells-1）/Redundancy查看cell软件版本imagehistoryimageinfo了解cell的温度dcli-gcell_group-lroot“ipmitoolsensor|grep‘InletAmbTemp'”cell存储节点的日志存放位置$ADR_BASE/diag/asm/cell/'hostname'/trace/alert.log$ADR_BASE/diag/asm/cell/'hostname'/trace/ms-odl.*$ADR_BASE/diag/asm/cell/'hostname'/trace/svtrc__0.trc—ps-ef|grep“cellsrv100"$ADR_BASE/diag/asm/cell/'hostname'/incident/*/var/log/messages*,dmesg/var/log/sa/*/var/log/cellos/*列出cell中的alerthistorylistalerthistorywherenotificationStatelike‘[]'andseveritylike‘[warning|critical]'andexaminedBy=NULL;为cell创建一个告警阈值cellclicreatethresholdCD_IO_ERRS_MINwarning=1,comparison='>=',occurrences=1,observation=1;cell可用性监控一般建议使用EMGCOracleExadataStorageServerManagementPlug-In监控如何禁用SmartScan?设置Cell_offload_processing=false如何禁用storageindex?设置_kcfis_storageidx_disabled=true如何禁用flashcache?11.2.0.2以后设置_kcfis_keep_in_cellfc_enabled=false11.2.0.1中设置_kcfis_control1=1cell相关的数据库视图有以下这些视图select*fromsys.GV_$CELL_STATE;select*fromsys.GV_$CELL;select*fromsys.GV_$CELL_THREAD_HISTORY;select*fromsys.GV_$CELL_REQUEST_TOTALS;select*fromsys.GV_$CELL_CONFIG;配置Inter-DatabaseIORMCellCLI>alteriormplan—dbplan=((name=production,level=1,allocation=100),(name=test,level=2,allocation=80),-(name=other,level=2,allocation=20))IORMPLANsuccessfullyalteredCellCLI>alteriormplanactiveIORMPLANsuccessfullyalteredCellCLI>listiormplandetailname:cell4_IORMPLANcatPlan:dbPlan:name=production,level=1,allocation=100name=test,level=2,allocation=80name=other,level=2,allocation=20status:active如何禁用布隆过滤BloomFliter设置_bloom_pruning_enabled=falseExadata数据备份backup备份速率Exadata下rman备份的速率从1通道到8通道大约为1003MB/s到2081MB/s，视乎配置不同也略微有区别recovery应用日志恢复速率exadatarecovery的速率大约为每秒600~1000MB/s的归档日志standbydatabase搭建对于50TB的standbydatabase搭建，若使用infiniband+4rman通道大约耗费5.5小时，若使用GigE则在18个小时左右Exadata恢复cell救护可以通过/opt/oracle.SupportTools/make_cellboot_usb脚本创建部USBcellboot_usb_in_rescure_modeExadata部署onecommand下载可以下载patch(9935478)ONECOMMANDFORExadata11gR2Exadata安装前准备工作下载安装介质包括Grid,Database,Patches等硬件设备到货验收并安装就绪规划DBM用的管理网，生产网，ILOM等用的网段和IP地址配置DNS服务器将IP地址和域名注册到DNS服务器配置NTP服务器网络连线环境检查检查DBM主机的eth0网卡是否可以通过cisco交换机被访问检查hardwareandfirmwareprofile是否正确验证InfiniBandNetwork验证网络连通性登陆第一台数据库服务器使用sh脚本验证网络连通性验证DNS是否正常验证NTP服务器是否正常安装ExadataStorageServerImagePatch(rootuser)在dbserver和cellserver上为root用户配置SSH#/opt/oracle.SupportTools/onecommand/setssh.sh-s-uroot-ppassword-nN-hdbs_group检查当前Cellstorageserver的ExadataImage版本安装最新的Patch具体步骤详见Readme验证当前ExadataImageversion#cd/opt/oracle.SupportTools/firstconf#dcli-lroot-gquarter‘imagehistory|grep-Version使用OneCommand工具完成DBM的配置安装#cd/opt/oracle.SupportTools/onecommandDisplaytheonecommandsteps#./deploy112.sh-i-Thestepsinorderare…Step0=ValidateThisNodeSetupStep1=SetupSSHForRootStep2=ValidateAllNodesStep3=UnzipFilesStep4=UpdateEtcHostsStep5=CreateCellipnitoraStep6=ValidateHWStep7=ValidateIBStep8=ValidateCellStep9=PingRdsCheckStep10=RunCalibrateStep11=ValidateTimeDateStep12=UpdateConfigStep13=CreateUserAccountsStep14=SetupSSHForUsersStep15=CreateOraHomesStep16=CreateGridDisksStep17=InstallGridSoftwareStep18=RunGridRootScriptsStep19=Install112DBSoftwareStep20=Create112ListenerStep21=RunAsmCaStep22=UnlockGIHomeStep23=UpdateOPatchStep24=ApplyBPStep25=RelinkRDSStep26=LockUpGIStep27=SetupCellEmailAlertsStep28=RunDbcaStep29=SetupEMDbControlStep30=ApplySecurityFixesStep31=ResecureMachineTorunacommand#./deploy112.sh-—NWhereNcorrespondstoastepnumberExampletorunstep0Exadata监控exachk健康检查脚本exachk脚本可以以daemon形式后台运行./exachk—start以daemon形式clustersupport运行./exachk-clusternodes[node1,[nodeN]]-dstart!Exadata文档信息Exadata的官方文档docs.oracle./cd/E50790_01/welcome.html另外文档还保存在您cell的/opt/oracle/cell/doc/下。Exadata硬件篇常规默认密码，以下是Exadata中cell/dbnodeIB等的默认密码:组件登陆默认密码StorageCellsrootnm2userwelcome1InfinibandSwitchrootnm2userwelcome1changemeDB节点rootwelcome1CELLCLIcelladminwelcomeILOMrootwelcome1KVMSwitchAdminornoneGigEswitch初始安装后asmsnmp的账号一般也是welcomel硬件常规巡检:在机房例行检查时，需要从Exadata机箱后方查看Exadata中是否有黄灯报警，如果有，记录位置，即时登录OEM/ILOM/集成的第三方监控工具查明原因，定位部件，即时维护。Exadata一体机健康检查脚本exachk，参考document1070954.1检测Exadata数据库机器上的硬件和固件版本是否匹配？/opt/oracle.SupportTools/CheckHWnFWProfile返回如下结果说明版本匹配：[SUCCESS]Thehardwareandfirmwareprofilematchesoneofthesupportedprofile检测软件版本与平台是否匹配？/opt/oracle.SupportTools/CheckSWProfile.sh-c为cell启用告警ALTERCELLsmtpServer='mailserver.maildomain.',-smtpFromAddr=firstname.lastnamemaildomain.',—smtpToAddr='firstname.lastnamemaildomain.',—smtpFrom='Exadatacell',-smtpPort='',—smtpUseSSL='TRUE',—notificationPolicy='critical,warning,clear',—notificationMethod='mail';altercellvalidatemail;监控磁盘故障当通过机房例行检查发现硬件黄灯警告或通过监控工具（命令行/ILOM/第三方工具）发现故障并确定位置后,可进行更换操作。更换StorageCell硬盘命令行登录Cell,判断故障硬盘,例如:CellCLI>LISTPHYSICALDISKWHEREdiskType=HardDiskANDstatus=criticalDETAIL观察DatabaseServer磁盘状态[rootdm01db01~]#cd/opt/MegaRAID/MegaCli/[rootdm01db01MegaCli]#./MegaCli64-Pdlist-aAll|grep“Slot\|Firmware”观察DatabaseServerRAID状态[rootdmOldbOlMegaCIi]#./MegaCli64-Ldlnfo-IAII-aAllStorageCell加电启动远程登陆StorageCell控制器lLOM,执行PowerOn,其它为系统的自动启动过程,知道StorageCell就绪CellCLl>LlSTGRlDDlSK若没有Active,需:CellCLl>ALTERGRlDDlSKALLACTlVE等griddiskActive后,ASM会自动同步，使griddiskOnline,查看状态：CellCLl>LlSTGRlDDlSKATTRlBUTESname,asmmodestatus确认ASM数据自动重新分布是否已经开始或完成。Grid用户登录+ASM实例执行：select*fromv$asm_operation;通过EM、SYSLOG、Cellcli、lLOM查看是否有告警解除信息检测memoryECC错误ipmitoolsellist|grepECC|cut-f1-d:|sort-u若发现Exadata上存在磁盘损毁则：使用/opt/oracle.SupportTools/sundiag.sh收集详细信息并发给oraclesupport检测cellserverCachePolicycell08#MegaCli64-LDInfo-Lall-aALL|grep'CurrentCachePolicy'CurrentCachePolicy:WriteThrough,ReadAheadNone,Direct,NoWriteCacheifBadBBUcell09#MegaCli64-LDInfo-Lall-aALL|grep'CurrentCachePolicy'CurrentCachePolicy:WriteBack,ReadAheadNone,Direct,NoWriteCacheifBadBBUDefaultCachePolicy:WriteBack,ReadAheadNone,Direct,NoWriteCacheifBadBBUCurrentCachePolicy:WriteThrough,ReadAheadNone,Direct,NoWriteCacheifBadBBUCachepolicyisinWBWouldrecommendproactivebatteryrepalcement.Example:a./opt/MegaRAID/MegaCli/MegaCli64-LDGetProp-Cache-LALL-aALL####(Willlistthecachepolicy)b./opt/MegaRAID/MegaCli/MegaCli64-LDSetProp-WB-LALL-aALL####(WilltrytochangetehpolicyfromxxtoWB)SopolicyChangetoWBwillnotcomeintoeffectimmediatelySetWritePolicytoWriteBackonAdapter0,VD0(targetid:0)successBatterycapacityisbelowthethresholdvalue检测cellBBU备用电池状态：cell08#/opt/MegaRAID/MegaCli/MegaCli64-AdpBbuCmd-GetBbuStatus-a0BBUstatusforAdapter:0BatteryType:iBBUVoltage:4061mVCurrent:0mATemperature:36CBBUFirmwareStatus:ChargingStatus:NoneVoltage:OKTemperature:OKLearnCycleRequested:NoLearnCycleActive:NoLearnCycleStatus:OKLearnCycleTimeout:NoI2cErrorsDetected:NoBatteryPackMissing:NoBatteryReplacementrequired:NoRemainingCapacityLow:YesPeriodicLearnRequired:NoBatterystate:GasGuageStatus:FullyDischarged:NoFullyCharged:YesDischarging:YesInitialized:YesRemainingTimeAlarm:NoRemainingCapacityAlarm:NoDischargeTerminated:NoOverTemperature:NoChargingTerminated:NoOverCharged:NoRelativeStateofCharge:99%ChargerSystemState:49168ChargerSystemCtrl:0Chargingcurrent:0mAAbsolutestateofcharge:21%MaxError:2%ExitCode:0x00批量检测BBU信息:dcli-g~/cell_group-lroot-t'{uname-srm;head-1/etc/*release;uptime|cut-d,-f1;imagehistory;ipmitoolsunoemcli"show/SPsystem_descriptionsystem_identifier"|grep=;ipmitoolsunoemcli"show/SP/policyFLASH_ACCELERATOR_CARD_INSTALLED/opt/MegaRAID/MegaCli/MegaCli64-AdpBbuCmd-GetBbuStatus-a0|egrep-i'BBU|Battery|Charge:|Fully|Low|Learn';}'|tee/tmp/ExaInfo.logExadata停机：确认无业务访问，以root用户登录第1个数据库服务器节点停止数据库（详见RAC/ASM维护之RAC启停章节）停止Cluster#GRID_HOME/grid/bin/crsctlstopcluster-all停除本机以外的数据库节点#dcli-lroot-cdm01db02,dm01db03,dm01db04shutdown-h-ynow停存储服务器cell_group可自编辑，执行时并可由root用户读取该文件（askmaclean.）另需参考StorageCell存储维护StorageCell停机章节信息后方可执行下述命令#dcli-lroot-gcell_groupshutdown-h-ynow停本机#shutdown-h-ynow此时可通过ILOM远程关机整机下电（关PDU）Exadata启动1、为机柜加电（SWITCH自然加电）打开PDU开关进行加电，服务器指示灯都变绿，慢闪若需手工开机数据库服务器、存储服务器需要按住其开关5秒。也可在ILOM中点击Cell的Poweron开关进行开机，服务器指示灯为绿色长亮，再点击DBServer的Poweron开关进行开机，服务器指示灯为绿色长亮。2、检查是否有黄灯报警。3、启动数据库、应用等。Infiniband篇启停IBSwitchInfiniBandSwitch电源的开启或关闭InfiniBandSwitch提供冗余电源，分别插在Exadata的2个冗余PDU电源上,并随PDU机柜电源开启或关闭若关闭InfiniBandSwitch需断掉InfiniBandSwitch的的冗余电源。2.查看OEM等是否有相关报警ILOM无法报警从cell1的cellcli中查看listalerthistory可以看到从db01查看网络拓扑状态[rootdm01db01~]#cd/opt/oracle.SupportTools/ibdiagtools[rootdm01db01ibdiagtools]#./verify-topology-thalfrack插入InfiniBand电源线査看InfiniBandSwitch正常启动检查IB链路状态#/opt/oracle.SupportTools/ibdiagtools/infinicheck-z#/opt/oracle.SupportTools/ibdiagtools/infinicheck查看IB网络拓扑状态登陆任意DatabaseServer，采用Exadata工具命令：[rootdm01db01~]#cd/opt/oracle.SupportTools/ibdiagtools[rootdm01db01ibdiagtools]#./verify-topology-thalfrack诊断IB链路没有错误#ibdiagnet-c1000-r查看IB网络连线以root用户登陆InfiniBandSwitchILOM，采用listlinkup命令显示：#listlinkupConnector0APresent<->I4Port31isip查看IB健康状态#showunhealthyOK-Nounhealthysensors.IB健康检查env_testIB故障处理1.确认已经备份IBSWITCH确认所有的cable已经label，之后从IBswitch上拔下cable拔下两根电源线poweroff取出IBswitch安装新IBswitch恢复IBswitch设置DisabletheSubnetManagerDisablesm连接cable确认cable连接的正确性/opt/oracle.SupportTools/ibdiagtools/verify-topology从任何主机上运行如下命令确认任何link没有错误ibdiagnet-c1000-EnabletheSubnetManagerusingEnablesmIB硬件监控showunhealthy&checkpowerSwitch端口错误ibqueryerrors.pl-sRcvSwRelayErrors,RcvRemotePhysErrors,XmtDiscards,XmtConstraintErrors,RcvConstraintErrors,ExcBufOverrunErrors,VL15DroppedLink状态HYPERLINK"file:///usr/sbin/iblinkinfo.pl"/usr/sbin/iblinkinfo.pl-RlSubnetmanager/usr/sbin/sminfoCISCO交换机例行维护操作采用CiscoIOS系统命令行方式,启动终端登陆管理网口IP:telnetxxx.xxx.xxx.xxx输入用户名(root)/口令(welcome1),进入enable模式:查看交换机的配置通过show命令查看:dm01sw-ip#showrunning-configBuildingconfiguration…显示信息包括交换机主机名称、IP地址、网关地址、IOS系统版本、时区信息、DNS配置、NTP配置、各网络端口配置、VLAN划分（全交换机一个VLAN）配置信息等。运行监控通过目前Cisco交换机监控的规进行监控。由于Cisco主要用于管理网使用,当完全不能访问时,只影响管理网的相关功能,不影响业务网的正常运行。当出现故障后,可采用目前Cisco交换机故障处理流程进行处理,并注意交换机主机名称、IP地址、网关地址、IOS系统版本、时区信息、DNS配置、NTP配置、各网络端口配置、VLAN划分（全交换机一个VLAN）等信息是否正确配置。KVM可通过OEMGC插件进行监控。PDU故障处理单路故障不影响Exadata的连续性运行，但需要即时报修更换（包括管理IP等）,以避免另外备份PDU也出现故障,导致Exadata非正常停机。