云原生Hiveonk8s环境部署

　　一、概述
　　Hive是基于Hadoop的一个数据仓库（DataAarehouse，简称数仓、DW），可以将结构化的数据文件映射为一张数据库表，并提供类SQL查询功能。是用于存储、分析、报告的数据系统。这里只讲部署，相关概念可以参考我这篇文章：大数据Hadoop之数据仓库Hive
　　Hive架构
　　Hive客户端架构
　　二、开始部署
　　因为hive依赖与Hadoop，所以这里是在把hive封装在Hadoophaonk8s编排中，关于更多，可以参考：【云原生】HadoopHAonk8s环境部署1）构建镜像
　　DockerfileFROMmyharbor。combigdatacentos：7。9。2009RUNrmfetclocaltimelnsvusrsharezoneinfoAsiaShanghaietclocaltimeechoAsiaShanghaietctimezoneRUNexportLANGzhCN。UTF8创建用户和用户组，跟yaml编排里的spec。template。spec。containers。securityContext。runAsUser：9999RUNgroupaddsystemgid9999adminuseraddsystemhomedirhomeadminuid9999gidadminadmin安装sudoRUNyumyinstallsudo；chmod640etcsudoers给admin添加sudo权限RUNechoadminALL（ALL）NOPASSWD：ALLetcsudoersRUNyumyinstallinstallnettoolstelnetwgetRUNmkdiroptapacheADDjdk8u212linuxx64。tar。gzoptapacheENVJAVAHOMEoptapachejdk1。8。0212ENVPATHJAVAHOMEbin：PATHENVHADOOPVERSION3。3。2ENVHADOOPHOMEoptapachehadoopENVHADOOPCOMMONHOME｛HADOOPHOME｝HADOOPHDFSHOME｛HADOOPHOME｝HADOOPMAPREDHOME｛HADOOPHOME｝HADOOPYARNHOME｛HADOOPHOME｝HADOOPCONFDIR｛HADOOPHOME｝etchadoopPATH｛PATH｝：｛HADOOPHOME｝binRUNcurlsilentoutputtmphadoop。tgzhttps：ftpstud。hsesslingen。depubMirrorsftp。apache。orgdisthadoopcommonhadoop｛HADOOPVERSION｝hadoop｛HADOOPVERSION｝。tar。gztardirectoryoptapachexzftmphadoop。tgzrmtmphadoop。tgzADDhadoop｛HADOOPVERSION｝。tar。gzoptapacheRUNlnsoptapachehadoop｛HADOOPVERSION｝｛HADOOPHOME｝ENVHIVEVERSION3。1。2ADDhive｛HIVEVERSION｝。tar。gzoptapacheENVHIVEHOMEoptapachehiveENVPATHHIVEHOMEbin：PATHRUNlnsoptapachehive｛HIVEVERSION｝｛HIVEHOME｝RUNchownRadmin：adminoptapacheWORKDIRoptapacheHdfsportsEXPOSE500105002050070500755009080209000MapredportsEXPOSE19888YarnportsEXPOSE8030803180328033804080428088OtherportsEXPOSE497072122
　　开始构建镜像dockerbuildtmyharbor。combigdatahadoophive：v3。3。23。1。2。nocache参数解释t：指定镜像名称。：当前目录Dockerfilef：指定Dockerfile路径nocache：不缓存dockerpushmyharbor。combigdatahadoophive：v3。3。23。1。22）添加Metastore服务编排1、配置
　　hadooptemplateshivehiveconfigmap。yamlapiVersion：v1kind：ConfigMapmetadata：name：｛｛includehadoop。fullname。｝｝hivelabels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝helm。shchart：｛｛includehadoop。chart。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝data：hivesite。xml：lt；？xmlversion1。0？lt；？xmlstylesheettypetextxslhrefconfiguration。xsl？configuration！配置hdfs存储目录propertynamehive。metastore。warehouse。dirnamevalueuserhiveremotewarehousevaluepropertypropertynamehive。metastore。localnamevaluefalsevalueproperty！所连接的MySQL数据库的地址，hivelocal是数据库，程序会自动创建，自定义就行propertynamejavax。jdo。option。ConnectionURLnamevaluejdbc：mysql：mysqlprimaryheadless。mysql：3306hivemetastore？createDatabaseIfNotExisttrueuseSSLfalseserverTimezoneAsiaShanghaivalueproperty！MySQL驱动propertynamejavax。jdo。option。ConnectionDriverNamenamevaluecom。mysql。cj。jdbc。Drivervalue！valuecom。mysql。jdbc。Drivervalueproperty！mysql连接用户propertynamejavax。jdo。option。ConnectionUserNamenamevaluerootvalueproperty！mysql连接密码propertynamejavax。jdo。option。ConnectionPasswordnamevalueWyfORdvwVmvalueproperty！元数据是否校验propertynamehive。metastore。schema。verificationnamevaluefalsevaluepropertypropertynamesystem：user。namenamevaluerootvaluedescriptionusernamedescriptionpropertypropertynamehive。metastore。urisnamevaluethrift：｛｛includehadoop。fullname。｝｝hivemetastore。｛｛。Release。Namespace｝｝。svc。cluster。local：9083valueproperty！hostpropertynamehive。server2。thrift。bind。hostnamevalue0。0。0。0valuedescriptionBindhostonwhichtoruntheHiveServer2Thriftservice。descriptionproperty！hs2端口默认是1000，为了区别，我这里不使用默认端口propertynamehive。server2。thrift。portnamevalue10000valuepropertypropertynamehive。server2。active。passive。ha。enablenamevaluetruevaluepropertyconfiguration2、控制器
　　hadooptemplateshivehiveserver2statefulset。yamlapiVersion：appsv1kind：StatefulSetmetadata：name：｛｛includehadoop。fullname。｝｝hivemetastoreannotations：checksumconfig：｛｛include（print。Template。BasePathhadoopconfigmap。yaml）。sha256sum｝｝labels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝helm。shchart：｛｛includehadoop。chart。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivemetastorespec：serviceName：｛｛includehadoop。fullname。｝｝hivemetastoreselector：matchLabels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivemetastorereplicas：｛｛。Values。hive。metastore。replicas｝｝template：metadata：labels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivemetastorespec：affinity：podAntiAffinity：｛｛ifeq。Values。antiAffinityhard｝｝requiredDuringSchedulingIgnoredDuringExecution：topologyKey：kubernetes。iohostnamelabelSelector：matchLabels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivemetastore｛｛elseifeq。Values。antiAffinitysoft｝｝preferredDuringSchedulingIgnoredDuringExecution：weight：5podAffinityTerm：topologyKey：kubernetes。iohostnamelabelSelector：matchLabels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivemetastore｛｛end｝｝terminationGracePeriodSeconds：0initContainers：name：waitnnimage：｛｛。Values。image。repository｝｝：｛｛。Values。image。tag｝｝command：〔sh，c，untilcurlm3sIhttp：｛｛includehadoop。fullname。｝｝hdfsnn｛｛sub。Values。hdfs。nameNode。replicas1｝｝。｛｛includehadoop。fullname。｝｝hdfsnn。｛｛。Release。Namespace｝｝。svc。cluster。local：9870egrepsilentHTTP1。1200OKHTTP1。1302Found；doechowaitingfornn；sleep1；done〕containers：name：hivemetastoreimage：｛｛。Values。image。repository｝｝：｛｛。Values。image。tag｝｝imagePullPolicy：｛｛。Values。image。pullPolicyquote｝｝command：binbashoptapachetmphadoopconfigbootstrap。shdresources：｛｛toYaml。Values。hive。metastore。resourcesindent10｝｝readinessProbe：tcpSocket：port：9083initialDelaySeconds：10timeoutSeconds：2livenessProbe：tcpSocket：port：9083initialDelaySeconds：10timeoutSeconds：2volumeMounts：name：hadoopconfigmountPath：optapachetmphadoopconfigname：hiveconfigmountPath：optapachehiveconfsecurityContext：runAsUser：｛｛。Values。securityContext。runAsUser｝｝privileged：｛｛。Values。securityContext。privileged｝｝volumes：name：hadoopconfigconfigMap：name：｛｛includehadoop。fullname。｝｝name：hiveconfigconfigMap：name：｛｛includehadoop。fullname。｝｝hive3、Service
　　hadooptemplateshivemetastoresvc。yamlAheadlessservicetocreateDNSrecordsapiVersion：v1kind：Servicemetadata：name：｛｛includehadoop。fullname。｝｝hivemetastorelabels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝helm。shchart：｛｛includehadoop。chart。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivemetastorespec：ports：name：metastoreport：｛｛。Values。service。hive。metastore。port｝｝nodePort：｛｛。Values。service。hive。metastore。nodePort｝｝type：｛｛。Values。service。hive。metastore。type｝｝selector：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivemetastore3）添加HiveServer2服务编排1、控制器
　　hadooptemplateshivehiveserver2statefulset。yamlapiVersion：appsv1kind：StatefulSetmetadata：name：｛｛includehadoop。fullname。｝｝hivehiveserver2annotations：checksumconfig：｛｛include（print。Template。BasePathhadoopconfigmap。yaml）。sha256sum｝｝labels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝helm。shchart：｛｛includehadoop。chart。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivehiveserver2spec：serviceName：｛｛includehadoop。fullname。｝｝hivehiveserver2selector：matchLabels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivehiveserver2replicas：｛｛。Values。hive。hiveserver2。replicas｝｝template：metadata：labels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivehiveserver2spec：affinity：podAntiAffinity：｛｛ifeq。Values。antiAffinityhard｝｝requiredDuringSchedulingIgnoredDuringExecution：topologyKey：kubernetes。iohostnamelabelSelector：matchLabels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivehiveserver2｛｛elseifeq。Values。antiAffinitysoft｝｝preferredDuringSchedulingIgnoredDuringExecution：weight：5podAffinityTerm：topologyKey：kubernetes。iohostnamelabelSelector：matchLabels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivehiveserver2｛｛end｝｝terminationGracePeriodSeconds：0initContainers：name：waithivemetastoreimage：｛｛。Values。image。repository｝｝：｛｛。Values。image。tag｝｝command：〔sh，c，until（echoq）telneteq｛｛includehadoop。fullname。｝｝hivemetastore。｛｛。Release。Namespace｝｝。svc。cluster。local｛｛。Values。service。hive。metastore。port｝｝devnull21；doechowaitingforhivemetastore；sleep1；done〕containers：name：hivehiveserver2image：｛｛。Values。image。repository｝｝：｛｛。Values。image。tag｝｝imagePullPolicy：｛｛。Values。image。pullPolicyquote｝｝command：binbashoptapachetmphadoopconfigbootstrap。shdresources：｛｛toYaml。Values。hive。metastore。resourcesindent10｝｝readinessProbe：tcpSocket：port：10000initialDelaySeconds：10timeoutSeconds：2livenessProbe：tcpSocket：port：10000initialDelaySeconds：10timeoutSeconds：2volumeMounts：name：hadoopconfigmountPath：optapachetmphadoopconfigname：hiveconfigmountPath：optapachehiveconfsecurityContext：runAsUser：｛｛。Values。securityContext。runAsUser｝｝privileged：｛｛。Values。securityContext。privileged｝｝volumes：name：hadoopconfigconfigMap：name：｛｛includehadoop。fullname。｝｝name：hiveconfigconfigMap：name：｛｛includehadoop。fullname。｝｝hive2、Service
　　hadooptemplateshivehiveserver2svc。yamlAheadlessservicetocreateDNSrecordsapiVersion：v1kind：Servicemetadata：name：｛｛includehadoop。fullname。｝｝hivehiveserver2labels：app。kubernetes。ioname：｛｛includehadoop。name。｝｝helm。shchart：｛｛includehadoop。chart。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivehiveserver2spec：ports：name：metastoreport：｛｛。Values。service。hive。hiveserver2。port｝｝nodePort：｛｛。Values。service。hive。hiveserver2。nodePort｝｝type：｛｛。Values。service。hive。hiveserver2。type｝｝selector：app。kubernetes。ioname：｛｛includehadoop。name。｝｝app。kubernetes。ioinstance：｛｛。Release。Name｝｝app。kubernetes。iocomponent：hivehiveserver24）修改values。yaml
　　hadoopvalues。yamlimage：repository：myharbor。combigdatahadoophivetag：v3。3。23。1。2pullPolicy：IfNotPresentTheversionofthehadooplibrariesbeingusedintheimage。hadoopVersion：3。3。2logLevel：INFOSelectantiAffinityaseitherhardorsoft，defaultissoftantiAffinity：softhdfs：nameNode：replicas：2pdbMinAvailable：1resources：requests：memory：256Micpu：10mlimits：memory：2048Micpu：1000mdataNode：Willbeusedasdfs。datanode。hostnameYoustillneedtosetupservicesingressforeveryDNDatanodeswillexpecttoexternalHostname：example。comexternalDataPortRangeStart：9866externalHTTPPortRangeStart：9864replicas：3pdbMinAvailable：1resources：requests：memory：256Micpu：10mlimits：memory：2048Micpu：1000mwebhdfs：enabled：truejounralNode：replicas：3pdbMinAvailable：1resources：requests：memory：256Micpu：10mlimits：memory：2048Micpu：1000mhive：metastore：replicas：1pdbMinAvailable：1resources：requests：memory：256Micpu：10mlimits：memory：2048Micpu：1000mhiveserver2：replicas：1pdbMinAvailable：1resources：requests：memory：256Micpu：10mlimits：memory：1024Micpu：500myarn：resourceManager：pdbMinAvailable：1replicas：2resources：requests：memory：256Micpu：10mlimits：memory：2048Micpu：2000mnodeManager：pdbMinAvailable：1ThenumberofYARNNodeManagerinstances。replicas：1Createstatefulsetsinparallel（K8S1。7）parallelCreate：falseCPUandmemoryresourcesallocatedtoeachnodemanagerpod。Thisshouldbetunedtofityourworkload。resources：requests：memory：256Micpu：500mlimits：memory：2048Micpu：1000mpersistence：nameNode：enabled：truestorageClass：hadoophannlocalstorageaccessMode：ReadWriteOncesize：1Gilocal：name：hadoophann0host：local168182110path：optbigdataservershadoophanndatadata1name：hadoophann1host：local168182111path：optbigdataservershadoophanndatadata1dataNode：enabled：trueenabledStorageClass：falsestorageClass：hadoophadnlocalstorageaccessMode：ReadWriteOncesize：1Gilocal：name：hadoophadn0host：local168182110path：optbigdataservershadoophadndatadata1name：hadoophadn1host：local168182110path：optbigdataservershadoophadndatadata2name：hadoophadn2host：local168182110path：optbigdataservershadoophadndatadata3name：hadoophadn3host：local168182111path：optbigdataservershadoophadndatadata1name：hadoophadn4host：local168182111path：optbigdataservershadoophadndatadata2name：hadoophadn5host：local168182111path：optbigdataservershadoophadndatadata3name：hadoophadn6host：local168182112path：optbigdataservershadoophadndatadata1name：hadoophadn7host：local168182112path：optbigdataservershadoophadndatadata2name：hadoophadn8host：local168182112path：optbigdataservershadoophadndatadata3volumes：name：dfs1mountPath：optapachehdfsdatanode1hostPath：optbigdataservershadoophadndatadata1name：dfs2mountPath：optapachehdfsdatanode2hostPath：optbigdataservershadoophadndatadata2name：dfs3mountPath：optapachehdfsdatanode3hostPath：optbigdataservershadoophadndatadata3journalNode：enabled：truestorageClass：hadoophajnlocalstorageaccessMode：ReadWriteOncesize：1Gilocal：name：hadoophajn0host：local168182110path：optbigdataservershadoophajndatadata1name：hadoophajn1host：local168182111path：optbigdataservershadoophajndatadata1name：hadoophajn2host：local168182112path：optbigdataservershadoophajndatadata1volumes：name：jnmountPath：optapachehdfsjournalnodeservice：nameNode：type：NodePortports：dfs：9000webhdfs：9870nodePorts：dfs：30900webhdfs：30870nameNode1：type：NodePortports：webhdfs：9870nodePorts：webhdfs：31870nameNode2：type：NodePortports：webhdfs：9870nodePorts：webhdfs：31871dataNode：type：NodePortports：webhdfs：9864nodePorts：webhdfs：30864resourceManager：type：NodePortports：web：8088nodePorts：web：30088resourceManager1：type：NodePortports：web：8088nodePorts：web：31088resourceManager2：type：NodePortports：web：8088nodePorts：web：31089journalNode：type：ClusterIPports：jn：8485nodePorts：jn：hive：metastore：type：NodePortport：9083nodePort：31183hiveserver2：type：NodePortport：10000nodePort：30000securityContext：runAsUser：9999privileged：true5）开始部署更新helmupgradehadoopha。hadoopnhadoopha重新安装helminstallhadoopha。hadoopnhadoophacreatenamespace更新helmupgradehadoopha。hadoopnhadoopha
　　NOTESNAME：hadoophaLASTDEPLOYED：ThuSep2923：42：022022NAMESPACE：hadoophaSTATUS：deployedREVISION：1TESTSUITE：NoneNOTES：1。YoucancheckthestatusofHDFSbyrunningthiscommand：kubectlexecnhadoophaithadoophahadoophdfsnn0opthadoopbinhdfsdfsadminreport2。Youcanlisttheyarnnodesbyrunningthiscommand：kubectlexecnhadoophaithadoophahadoopyarnrm0opthadoopbinyarnnodelist3。CreateaportforwardtotheyarnresourcemanagerUI：kubectlportforwardnhadoophahadoophahadoopyarnrm08088：8088Thenopentheuiinyourbrowser：openhttp：localhost：80884。Youcanrunincludedhadooptestslikethis：kubectlexecnhadoophaithadoophahadoopyarnnm0opthadoopbinhadoopjaropthadoopsharehadoopmapreducehadoopmapreduceclientjobclient3。3。2tests。jarTestDFSIOwritenrFiles5fileSize128MBresFiletmpTestDFSIOwrite。txt5。Youcanlistthemapreducejobslikethis：kubectlexecnhadoophaithadoophahadoopyarnrm0opthadoopbinmapredjoblist6。Thischartcanalsobeusedwiththezeppelincharthelminstallnamespacehadoophasethadoop。useConfigMaptrue，hadoop。configMapNamehadoophahadoopstablezeppelin7。Youcanscalethenumberofyarnnodeslikethis：helmupgradehadoophasetyarn。nodeManager。replicas4stablehadoopMakesuretoupdatethevalues。yamlifyouwanttomakethispermanent。
　　6）测试验证
　　查看kubectlgetpods，svcnhadoophaowide
　　测试beelineujdbc：hive2：localhost：10000nadmincreatedatabasetest；CREATETABLEIFNOTEXISTStest。person1（idINTCOMMENTID，nameSTRINGCOMMENT名字，ageINTCOMMENT年龄，likesARRAYSTRINGCOMMENT爱好，addressMAPSTRING，STRINGCOMMENT地址）ROWFORMATDELIMITEDFIELDSTERMINATEDBY，COLLECTIONITEMSTERMINATEDBYMAPKEYSTERMINATEDBY：LINESTERMINATEDBY；
　　7）卸载helmuninstallhadoophanhadoophakubectldeletepodnhadoophakubectlgetpodnhadoophaawkNR1｛print1｝forcekubectlpatchnshadoophap｛metadata：｛finalizers：null｝｝kubectldeletenshadoophaforcermfroptbigdataservershadoopha｛nn，dn，jn｝datadata｛1。。3｝
　　git下载地址：https：gitee。comhadoopbigdatahadoophaonk8s
　　这里只是把hive相关的部分编排列出来了，有疑问的小伙伴欢迎给我留言，对应的修改也提交到git上了，有需要的小伙伴自行下载，hive的编排部署就先到这里了，后续会持续分享【云原生大数据】相关的教程，请小伙伴耐心等待