Redis简介Redis全称REmoteDIctionaryServer，是一种开源的内存数据存储，通常用作数据库、缓存或消息代理。它可以存储和操作高级数据类型，例如列表、映射、集合和排序集合。由于Redis接受多种格式的密钥，因此可以在服务器端进行操作，减少客户端的工作量。它仅将磁盘用于持久性，将数据库完全保存在内存中。Redis是一种流行的数据存储解决方案，被GitHub、Pinterest、Snapchat、Twitter、StackOverflow、Flickr等科技巨头使用。为什么使用Redis它非常快。它是用ANSIC编写的，可在POSIX系统（例如Linux、MacOSX和Solaris）上运行。Redis通常被列为最流行的键/值数据库和最流行的用于容器的NoSQL数据库。它的缓存解决方案减少了对云数据库后端的调用次数。应用程序可以通过其客户端API库访问它。所有流行的编程语言都支持Redis。它是开源且稳定的。什么是Redis集群Redis集群是一组Redis实例，旨在通过对数据库进行分区来扩展数据库，使其更具弹性。集群的每个成员（无论是主要成员还是次要成员）都管理一个哈希槽子集。如果master变得不可访问，它的slaves将被提升为master。在一个由三个主节点组成的最小Redis集群中，每个主节点有一个从节点（用于最小故障转移），每个主节点都分配了一个范围在0到16,383之间的哈希槽。节点A包含从0到5000的哈希槽，节点B从5001到10000，节点C从10001到16383。集群内的通信是通过内部总线，使用协议来传播有关集群的信息或发现新节点。在Kubernetes中部署Redis集群在Kubernetes中部署Redis集群具有挑战性，因为每个Redis实例都依赖一个配置文件来跟踪其他集群实例及其角色。为此，我们需要结合使用KubernetesStatefulSets和PersistentVolumes。克隆部档案gitclonehttps://github.com/llmgo/redis-sts.git创建statefulset类型资源[root@node01redis-sts]#catredis-sts.yml---apiVersion:v1kind:ConfigMapmetadata:name:redis-clusterdata:update-node.sh:|#!/bin/shREDIS_NODES="/data/nodes.conf"sed-i-e"/myself/s/[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}/${POD_IP}/"${REDIS_NODES}exec"$@“redis.conf:|+cluster-enabledyescluster-require-full-coveragenocluster-node-timeout15000cluster-config-file/data/nodes.confcluster-migration-barrier1appendonlyyesprotected-modeno---apiVersion:apps/v1kind:StatefulSetmetadata:name:redis-clusterspec:serviceName:redis-clusterreplicas:6selector:matchLabels:app:redis-clustertemplate:metadata:labels:app:redis-clusterspec:containers:-name:redisimage:redis:5.0.5-alpineports:-containerPort:6379name:client-containerPort:16379name:gossipcommand:["/conf/update-node.sh","re??dis-server","/conf/redis.conf"]env:-name:POD_IPvalueFrom:fieldRef:fieldPath:status.podIPvolumeMounts:-name:confmountPath:/confreadOnly:false-name:datamountPath:/datareadOnly:falsevolumes:-name:confconfigMap:name:redis-clusterdefaultMode:0755volumeClaimTemplates:-metadata:name:dataspec:accessModes：[“ReadWriteOnce”]资源：请求：存储：5GistorageClassName：标准$kubectlapply-fredis-sts.ymlconfigmap/redis-clustercreatedstatefulset.apps/redis-clustercreated$kubectlgetpods-lapp=redis-clusterNAMEREADYSTATUSRESTARTSAGEredis-cluster-01/1Running053-redis-cluster11/1Running049sredis-cluster-21/1Running046sredis-cluster-31/1Running042sredis-cluster-41/1Running038sredis-cluster-51/1Running034s创建服务[root@node01redis-sts]#catredis-svc.yml---apiVersion:v1kind:name:redis-clusterspec:type:ClusterIPclusterIP:10.96.0.100ports:-port:6379targetPort:6379name:client-port:16379targetPort:16379name:gossipselector:app:redis-cluster$kubectlapply-fredis-svc.ymlservice/redis-clustercreated$kubectlgetsvcredis-clusterNAMETYPECLUSTER-IPEXTERNAL-IPPORT(S)AGEredis-clusterClusterIP10.96.0.1006379/TCP,16379/TCP35s初始化redis集群下一步是形成Redis集群为此，我们运行以下命令并输入是接受配置。前三个节点成为主节点，后三个节点成为从节点。$kubectlexec-itredis-cluster-0--redis-cli--clustercreate--cluster-replicas1$(kubectlgetpods-lapp=redis-cluster-ojsonpath='{range.items[*]}{.status.podIP}:6379')>>>Performinghashslotsallocationon6nodes...Master[0]->Slots0-5460Master[1]->Slots5461-10922Master[2]->Slots10923-16383Addingreplica10.244.2.11:6379to10.244.9.19:6379Addingreplica10.2044.9:6379to10.244.6.10:6379Addingreplica10.244.8.15:6379to10.244.7.8:6379M:00721c43db194c8f2cacbafd01fd2be6a2fede2810.244.9.19:6379slots:[0-5460](5461slots)masterM:9c36053912dec8cb20a599bda202a654f241484f10.244.6.10:6379slots:[5461-10922](5462slots)masterM:2850f24ea6367de58fb50e632fc56fe4ba5ef01610.244.7.8:6379slots:[10923-16383](5461slots)masterS:554a58762e3dce23ca5a75886d0ccebd2d58250210.244.8.15:6379replicates2850f24ea6367de58fb50e632fc56fe4ba5ef016S:20028fd0b79045489824eda71fac9898f17af89610.244.2.11:6379replicates00721c43db194c8f2cacbafd01fd2be6a2fede28S:87e8987e314e4e5d4736e5818651abc1ed6ddcd910.244.9.20:6379replicates9c36053912dec8cb20a599bda202a654f241484fCanIsettheaboveconfiguration?(type'yes'toaccept):yes>>>Nodesconfigurationupdated>>>Assignadifferentconfigepochtoeachnode>>>SendingCLUSTERMEETmessagestojointheclusterWaitingfortheclustertojoin...>>>PerformingClusterCheck(usingnode10.244.9.19:6379)M:00721c43db194c8f2cacbafd01fd2be6a2fede2810.244.9.19:6379slots:[0-5460](5461slots)master1additionalreplica(s)M:9c36053912dec8cb20a599bda202a654f241484f10.244.6.10:6379slots:[5461-10922](5462slots)master1additionalreplica(s)S:87e8987e314e4e5d4736e5818651abc1ed6ddcd910.244.9.20:6379slots:(0slots)slavereplicates9c36053912dec8cb20a599bda202a654f241484fS:554a58762e3dce23ca5a75886d0ccebd2d58250210.244.8.15:6379slots:(0slots)slavereplicates2850f24ea6367de58fb50e632fc56fe4ba5ef016S:20028fd0b79045489824eda71fac9898f17af89610.244.2.11:6379slots:(0slots)slavereplicates00721c43db194c8f2cacbafd01fd2be6a2fede28M:2850f24ea6367de58fb50e632fc56fe4ba5ef01610.244.7.8:6379slots:[10923-16383](5461slots)master1additionalreplica(s)[OK]Allnodesagreeaboutslotsconfiguration.>>>Checkforopenslots...>>>Checkslotscoverage...[OK]All16384slotscovered.验证集群[root@node01redis-sts]#kubectlexec-itredis-cluster-0--redis-cliclusterinfocluster_state:okcluster_slots_assigned:16384cluster_slots_ok:16384cluster_slots_pfail:0cluster_slots_fail:0cluster_known_nodes:6cluster_size:3cluster_current_epoch:6cluster_my_epoch:1cluster_stats_messages_ping_sent:16cluster_stats_messages_pong_sent:22cluster_stats_messages_sent:38cluster_stats_messages_ping_received:17cluster_stats_messages_pong_received:16cluster_stats_messages_meet_received:5cluster_stats_messages_received:38[root@node01redis-sts]#forxin$(seq05);doecho"redis-cluster-$x";kubectlexecredis-cluster-$x--redis-clirole;echo;doneredis-cluster-0master1410.244.2.11637914redis-cluster-1master2810.244.9.20637928redis-cluster-2master2810.244.8.15637928redis-cluster-3slave10.244.7.86379connected28redis-cluster-4slave10.244.9.196379connected14redis-cluster-5slave10.244.6.106379connected28我们要用模拟节点'testcluster来模拟之前的集群我们将部署一个简单的Python应用程序，对于后者，我们将删除一个节点并观察集群行为。部署点击计数器应用程序我们将向集群部署一个简单的应用程序，并在其前面放置一个负载均衡器。此应用程序的目的是在将计数器值作为HTTP响应返回之前递增计数器并将其存储在Redis集群中。$kubectlapply-fapp-deployment-service.ymlservice/hit-counter-lbcreateddeployment.apps/hit-counter-appcreated在此过程中，如果我们继续加载页面，计数器会不断增加，删除pod后，我们看到没有数据丢失。$curl`kubectlgetsvchit-counter-lb-ojson|jq-r.spec.clusterIP`Ihavebeenhit20timessincedeployment.$curl`kubectlgetsvchit-counter-lb-ojson|jq-r.spec.clusterIP`Ihavebeenhit21timessincedeployment.$curl`kubectlgetsvchit-counter-lb-ojson|jq-r.spec.clusterIP`Ihavebeenhit22timessincedeployment.$kubectldeletepodsredis-cluster-0pod"redis-cluster-0"删除$kubectldeletepodsredis-cluster-1pod"redis-cluster-1"删除$curl`kubectlgetsvchit-counter-lb-ojson|jq-r.spec.clusterIP`自部署以来我已经命中23次。
