scofield 菜鸟运维杂谈
方案一 自定义yaml文件安装redis cluster背景在Kubernetes中部署Redis集群面临挑战，因为每个Redis实例都依赖于一个配置文件，该文件可以跟踪其他集群实例及其角色。为此，需要结合使用Kubernetes StatefulSets和PersistentVolumes实现。
redis cluster架构示意图：
---apiVersion: v1kind: ConfigMapmetadata: name: redis-clusterdata: update.sh: | #!/bin/sh REDIS_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: |+ bind 0.0.0.0 cluster-enabled yes cluster-require-full-coverage no cluster-node-timeout 30000 cluster-config-file /data/nodes.conf cluster-migration-barrier 1 appendonly yes protected-mode no---apiVersion: apps.kruise.io/v1beta1# apiVersion: apps/v1kind: StatefulSetmetadata: name: redis-clusterspec: serviceName: redis-cluster replicas: 6 selector: matchLabels: app: redis-cluster template: metadata: labels: app: redis-cluster spec: containers: - name: redis image: redis:6.2.1-alpine ports: - containerPort: 6379 name: client - containerPort: 16379 name: gossip command: ["/conf/update.sh", "redis-server", "/conf/redis.conf"] env: - name: POD_IP valueFrom: fieldRef: fieldPath: status.podIP volumeMounts: - name: conf mountPath: /conf readOnly: false - name: data mountPath: /data readOnly: false volumes: - name: conf configMap: name: redis-cluster defaultMode: 0755 volumeClaimTemplates: - metadata: name: data spec: accessModes: [ "ReadWriteOnce" ] resources: requests: storage: 100Gi storageClassName: rbd注释：
clu```
ster-migration-barrier 那些分配后仍然剩余migration barrier个从节点的主节点才会触发节点分配，而不是分配前有migration barrier个从节点的主节点就会触发节点分配，默认是1，生产环境建议维持默认值
protected-mode no 参数是为了禁止外网访问redis，如果启用了，则只能够通过lookback ip（127.0.0.1）访问Redis，如果外网访问redis，会报出异常。
apiVersion: apps.kruise.io/v1beta1 控制器这里使用kruise提供的Advanced StatefulSet，如果集群没有安装kruise，可以使用 apps/v1
### 安装redis-cluster-----[root@qd01-stop-k8s-master001 redis]# kubectl apply -f install-redis.yaml
configmap/redis-cluster created
statefulset.apps.kruise.io/redis-cluster created
[root@qd01-stop-k8s-master001 redis]# kubectl get po -n op
NAME READY STATUS RESTARTS AGE
redis-cluster-0 1/1 Running 0 3m26s
redis-cluster-1 1/1 Running 0 3m14s
redis-cluster-2 1/1 Running 0 2m54s
redis-cluster-3 1/1 Running 0 2m23s
redis-cluster-4 1/1 Running 0 2m14s
redis-cluster-5 1/1 Running 0 114s
### 创建redis-cluster service-----apiVersion: v1
kind: Service
metadata:
name: redis-cluster
namespace: op
spec:
type: ClusterIP
ports:
- port: 6379
targetPort: 6379
name: client - port: 16379
targetPort: 16379
name: gossip
selector:
app: redis-cluster
[root@qd01-stop-k8s-master001 redis]# kubectl apply -f redis-svc.yml
service/redis-cluster created
[root@qd01-stop-k8s-master001 redis]# kubectl get svc -n op
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
redis-cluster ClusterIP 10.97.197.224 <none> 6379/TCP,16379/TCP 9s
测试能后连通
[root@qd01-stop-k8s-master001 redis]# telnet 10.97.197.224 6379
Trying 10.97.197.224...
Connected to 10.97.197.224.
Escape character is '^]'.
### 初始化redis-cluster-----执行如下命令，获取到pod IP，然后使用redis-cli --cluster创建集群[root@qd01-stop-k8s-master001 redis]# kubectl -n op exec -it redis-cluster-0 -- redis-cli --cluster create --cluster-replicas 1 $(kubectl -n op get pods -l app=redis-cluster -o jsonpath='{range.items[*]}{.status.podIP}:6379 {end}')
Performing hash slots allocation on 6 nodes...
Master[0] -> Slots 0 - 5460
Master[1] -> Slots 5461 - 10922
Master[2] -> Slots 10923 - 16383
Adding replica 100.88.43.67:6379 to 100.64.147.152:6379
Adding replica 100.113.170.5:6379 to 100.98.174.217:6379
Adding replica 100.64.147.153:6379 to 100.80.158.227:6379
M: b47b27a3dbddf3fc1370cbe14ae753f4fce20b04 100.64.147.152:6379
slots:[0-5460] (5461 slots) master
M: 09543217c903350e963fc4fdf4acb73f8a1b7f8b 100.98.174.217:6379
slots:[5461-10922] (5462 slots) master
M: 5389ace495b68eeac85370d6783648dff68f2fb6 100.80.158.227:6379
slots:[10923-16383] (5461 slots) master
S: b1f39714c006ae55b12b18e6537303d7a00e1704 100.64.147.153:6379
replicates 5389ace495b68eeac85370d6783648dff68f2fb6
S: 0113f4668ec2f3ca2e9470c44bd5faab532b0936 100.88.43.67:6379
replicates b47b27a3dbddf3fc1370cbe14ae753f4fce20b04
S: e1e2f18ae66c79f1943390beabb59613abbad38a 100.113.170.5:6379
replicates 09543217c903350e963fc4fdf4acb73f8a1b7f8b
Can I set the above configuration? (type 'yes' to accept): yes
Nodes configuration updated
Assign a different config epoch to each node
Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join
..
Performing Cluster Check (using node 100.64.147.152:6379)
M: b47b27a3dbddf3fc1370cbe14ae753f4fce20b04 100.64.147.152:6379
slots:[0-5460] (5461 slots) master
1 additional replica(s)
S: 0113f4668ec2f3ca2e9470c44bd5faab532b0936 100.88.43.67:6379
slots: (0 slots) slave
replicates b47b27a3dbddf3fc1370cbe14ae753f4fce20b04
M: 09543217c903350e963fc4fdf4acb73f8a1b7f8b 100.98.174.217:6379
slots:[5461-10922] (5462 slots) master
1 additional replica(s)
M: 5389ace495b68eeac85370d6783648dff68f2fb6 100.80.158.227:6379
slots:[10923-16383] (5461 slots) master
1 additional replica(s)
S: e1e2f18ae66c79f1943390beabb59613abbad38a 100.113.170.5:6379
slots: (0 slots) slave
replicates 09543217c903350e963fc4fdf4acb73f8a1b7f8b
S: b1f39714c006ae55b12b18e6537303d7a00e1704 100.64.147.153:6379
slots: (0 slots) slave
replicates 5389ace495b68eeac85370d6783648dff68f2fb6
[OK] All nodes agree about slots configuration.
Check for open slots...
Check slots coverage...
[OK] All 16384 slots covered.### 验证集群信息
[root@qd01-stop-k8s-master001 redis]# kubectl -n op exec -it redis-cluster-0 -- redis-cli cluster infocluster_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:178cluster_stats_messages_pong_sent:181cluster_stats_messages_sent:359cluster_stats_messages_ping_received:176cluster_stats_messages_pong_received:178cluster_stats_messages_meet_received:5cluster_stats_messages_received:359[root@qd01-stop-k8s-master001 redis]# kubectl -n op exec -it redis-cluster-0 -- redis-cli cluster nodes0113f4668ec2f3ca2e9470c44bd5faab532b0936 100.88.43.67:6379@16379 slave b47b27a3dbddf3fc1370cbe14ae753f4fce20b04 0 1615348311156 1 connected09543217c903350e963fc4fdf4acb73f8a1b7f8b 100.98.174.217:6379@16379 master - 0 1615348314162 2 connected 5461-10922b47b27a3dbddf3fc1370cbe14ae753f4fce20b04 100.64.147.152:6379@16379 myself,master - 0 1615348312000 1 connected 0-54605389ace495b68eeac85370d6783648dff68f2fb6 100.80.158.227:6379@16379 master - 0 1615348312000 3 connected 10923-16383e1e2f18ae66c79f1943390beabb59613abbad38a 100.113.170.5:6379@16379 slave 09543217c903350e963fc4fdf4acb73f8a1b7f8b 0 1615348313160 2 connectedb1f39714c006ae55b12b18e6537303d7a00e1704 100.64.147.153:6379@16379 slave 5389ace495b68eeac85370d6783648dff68f2fb6 0 1615348312158 3 connected从输出可以看到，集群总共6个节点，三主三从
方案二 使用kubeDB安装redis安装kubeDB1、安装KubeDB
获取AppsCode License https://license-issuer.appscode.com/
下载KubeDB charts https://github.com/appscode/charts/tree/master/stable/kubedb-community
[root@qd01-stop-k8s-master001 kubedb-community]# unzip kubedb-community-v0.16.2.tgz[root@qd01-stop-k8s-master001 kubedb-community]# cd kubedb-community[root@qd01-stop-k8s-master001 kubedb-community]# ls -altotal 96drwxr-xr-x 4 root root 158 Mar 10 15:26 .drwxr-xr-x 3 root root 66 Mar 10 15:24 ..-rw-r--r-- 1 root root 351 Feb 16 09:55 Chart.yamldrwxr-xr-x 2 root root 28 Mar 10 15:24 ci-rw-r--r-- 1 root root 493 Feb 16 09:55 doc.yaml-rw-r--r-- 1 root root 353 Feb 16 09:55 .helmignore-rw-r--r-- 1 root root 24422 Feb 16 09:55 README.mddrwxr-xr-x 2 root root 4096 Mar 10 15:24 templates-rw-r--r-- 1 root root 47437 Feb 16 09:55 values.openapiv3_schema.yaml-rw-r--r-- 1 root root 5230 Feb 16 09:55 values.yaml修改values.yaml，把License文件放到kubedb-community目录下
2、使用helm安装
[root@qd01-stop-k8s-master001 kubedb-community]# helm install kubedb-community --namespace kube-system --set-file license=./kubedb-community-license.txt -f values.yaml .NAME: kubedb-communityLAST DEPLOYED: Wed Mar 10 15:38:59 2021NAMESPACE: kube-systemSTATUS: deployedREVISION: 1TEST SUITE: NoneNOTES:To verify that KubeDB has started, run: kubectl get deployment --namespace kube-system -l "app.kubernetes.io/name=kubedb-community,app.kubernetes.io/instance=kubedb-community"Now install/upgrade appscode/kubedb-catalog chart.To install, run: helm install kubedb-catalog appscode/kubedb-catalog --version v0.16.2 --namespace kube-systemTo upgrade, run: helm upgrade kubedb-catalog appscode/kubedb-catalog --version v0.16.2 --namespace kube-system执行如下命令查看是否安装完成[root@qd01-stop-k8s-master001 kubedb-community]# kubectl get deployment --namespace kube-system -l "app.kubernetes.io/name=kubedb-community,app.kubernetes.io/instance=kubedb-community"NAME READY UP-TO-DATE AVAILABLE AGEkubedb-community 1/1 1 1 38s等待crds注册成功[root@qd01-stop-k8s-master001 kubedb-community]# kubectl get crds -l app.kubernetes.io/name=kubedb -wNAME CREATED ATelasticsearches.kubedb.com 2021-03-10T07:39:42Zelasticsearchversions.catalog.kubedb.com 2021-03-10T07:39:45Zetcds.kubedb.com 2021-03-10T07:39:42Zetcdversions.catalog.kubedb.com 2021-03-10T07:39:45Zmemcacheds.kubedb.com 2021-03-10T07:39:43Zmemcachedversions.catalog.kubedb.com 2021-03-10T07:39:45Zmongodbs.kubedb.com 2021-03-10T07:39:43Zmongodbversions.catalog.kubedb.com 2021-03-10T07:39:45Zmysqls.kubedb.com 2021-03-10T07:39:43Zmysqlversions.catalog.kubedb.com 2021-03-10T07:39:46Zperconaxtradbs.kubedb.com 2021-03-10T07:39:43Zperconaxtradbversions.catalog.kubedb.com 2021-03-10T07:39:46Zpgbouncers.kubedb.com 2021-03-10T07:39:44Zpgbouncerversions.catalog.kubedb.com 2021-03-10T07:39:46Zpostgreses.kubedb.com 2021-03-10T07:39:44Zpostgresversions.catalog.kubedb.com 2021-03-10T07:39:46Zproxysqls.kubedb.com 2021-03-10T07:39:44Zproxysqlversions.catalog.kubedb.com 2021-03-10T07:39:46Zredises.kubedb.com 2021-03-10T07:39:45Zredisversions.catalog.kubedb.com 2021-03-10T07:39:46Z3、安装KubeDB Catalog
同样，先下载https://github.com/appscode/charts/tree/master/stable/kubedb-catalog
[root@qd01-stop-k8s-master001 kubedb-catalog]# tar -zxf kubedb-catalog-v0.16.2.tgz[root@qd01-stop-k8s-master001 kubedb-catalog]# cd kubedb-catalog[root@qd01-stop-k8s-master001 kubedb-catalog]# ls -altotal 24drwxr-xr-x 3 root root 148 Mar 10 15:48 .drwxr-xr-x 3 root root 28 Mar 10 15:48 ..-rw-r--r-- 1 root root 321 Jan 26 20:08 Chart.yaml-rw-r--r-- 1 root root 467 Jan 26 20:08 doc.yaml-rw-r--r-- 1 root root 353 Jan 26 20:08 .helmignore-rw-r--r-- 1 root root 3195 Jan 26 20:08 README.mddrwxr-xr-x 12 root root 188 Mar 10 15:48 templates-rw-r--r-- 1 root root 744 Jan 26 20:08 values.openapiv3_schema.yaml-rw-r--r-- 1 root root 1070 Jan 26 20:08 values.yaml[root@qd01-stop-k8s-master001 kubedb-catalog]# helm install kubedb-catalog --namespace kube-system -f values.yaml .NAME: kubedb-catalogLAST DEPLOYED: Wed Mar 10 15:50:50 2021NAMESPACE: kube-systemSTATUS: deployedREVISION: 1TEST SUITE: None1、先看下官方给的redis生命周期示意图
kubedb安装redis支持如下特性
Features AvailabilityClustering ✓Instant Backup ✗Scheduled Backup ✗Persistent Volume ✓Initialize using Snapshot ✗Initialize using Script ✗Custom Configuration ✓Using Custom docker image ✓Builtin Prometheus Discovery ✓Using Prometheus operator ✓2、查看支持的版本
[root@qd01-stop-k8s-master001 kubedb-catalog]# kubectl get redisversionsNAME VERSION DB_IMAGE DEPRECATED AGE4.0.11 4.0.11 kubedb/redis:4.0.11 15m4.0.6-v2 4.0.6 kubedb/redis:4.0.6-v2 15m5.0.3-v1 5.0.3 kubedb/redis:5.0.3-v1 15m6.0.6 6.0.6 kubedb/redis:6.0.6 15m3、编辑yaml安装文件
可以参照https://github.com/kubedb/docs/blob/v2021.01.26/docs/examples/redis/clustering/demo-1.yaml
这里选择安装6.0.6这个版本，我的集群storageClassName: "rbd"，请根据实际修改
如果想自定义redis.conf,请参考https://github.com/kubedb/docs/blob/v2021.01.26/docs/examples/redis/custom-config/redis-custom.yaml
apiVersion: kubedb.com/v1alpha2kind: Redismetadata: name: redis-cluster namespace: opspec: version: 6.0.6 mode: Cluster cluster: master: 3 replicas: 1 storageType: Durable storage: resources: requests: storage: 1Gi storageClassName: "rbd" accessModes: - ReadWriteOnce执行安装
[root@qd01-stop-k8s-master001 kubedb-community]# kubectl apply -f redis-cluster.yamlredis.kubedb.com/redis-cluster created安装完成，可以如下查看[root@qd01-stop-k8s-master001 kubedb-community]# kubectl get rd,po -n opNAME VERSION STATUS AGEredis.kubedb.com/redis-cluster 6.0.6 Provisioning 6m55sNAME READY STATUS RESTARTS AGEpod/redis-cluster-shard0-0 1/1 Running 0 6m54spod/redis-cluster-shard0-1 1/1 Running 0 6m18spod/redis-cluster-shard1-0 1/1 Running 0 5m38spod/redis-cluster-shard1-1 1/1 Running 0 5m1spod/redis-cluster-shard2-0 1/1 Running 0 4m30spod/redis-cluster-shard2-1 1/1 Running 0 4m8s[root@qd01-stop-k8s-master001 redis]# kubectl get svc -n opNAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGEredis-cluster ClusterIP 10.97.197.224 <none> 6379/TCP 5h16mredis-cluster-pods ClusterIP None <none> 6379/TCP 17m4、验证集群
[root@qd01-stop-k8s-master001 kubedb-community]# kubectl get pods -n op -o jsonpath='{range.items[*]}{.metadata.name} ---------- {.status.podIP}:6379{"\t\n"}{end}' | grep redisredis-cluster-shard0-0 ---------- 100.64.147.156:6379redis-cluster-shard0-1 ---------- 100.98.174.218:6379redis-cluster-shard1-0 ---------- 100.126.252.204:6379redis-cluster-shard1-1 ---------- 100.113.170.6:6379redis-cluster-shard2-0 ---------- 100.107.55.69:6379redis-cluster-shard2-1 ---------- 100.78.230.4:6379[root@qd01-stop-k8s-master001 redis]# kubectl -n op exec -it redis-cluster-shard0-0 -- redis-cli cluster infocluster_state:okcluster_slots_assigned:16384cluster_slots_ok:16384cluster_slots_pfail:0cluster_slots_fail:0cluster_known_nodes:6cluster_size:3cluster_current_epoch:3cluster_my_epoch:1cluster_stats_messages_ping_sent:864cluster_stats_messages_pong_sent:882cluster_stats_messages_sent:1746cluster_stats_messages_ping_received:879cluster_stats_messages_pong_received:864cluster_stats_messages_meet_received:3cluster_stats_messages_received:1746[root@qd01-stop-k8s-master001 redis]# kubectl -n op exec -it redis-cluster-shard0-0 -- redis-cli cluster nodes1895cb4b9c31b848666c61000e502f55a29a8255 100.64.147.155:6379@16379 master - 0 1615365162008 2 connected 5461-1092230bdbf2ca37001774498a9b935afbc1cd2ce389c 100.126.252.203:6379@16379 slave 2c06092fafa99e0158e39e6237a04fed25be3550 0 1615365163000 1 connected9b2cfbd5c1b417121d410141b6da9512ad29ce3c 100.78.230.3:6379@16379 slave e83446c368839c5fdccf5f70e3b1004eb67cb651 0 1615365163512 3 connected2c06092fafa99e0158e39e6237a04fed25be3550 100.82.197.130:6379@16379 myself,master - 0 1615365162000 1 connected 0-54601379d2b20f26ab13d53068d276ec5d988b7a0273 100.64.122.197:6379@16379 slave 1895cb4b9c31b848666c61000e502f55a29a8255 0 1615365163000 2 connectede83446c368839c5fdccf5f70e3b1004eb67cb651 100.107.55.68:6379@16379 master - 0 1615365164014 3 connected 10923-16383PS：文章会同步到dev.kubeops.net
下面的是我的公众号二维码，欢迎关注。