k8s安装kafka

安装kafka

安装zookeeper

安装zookeeper可以采用k8s的StatefulSet方式,非常方便,在安装kafka过程中遇到很多问题,但是zookeeper非常文档

apiVersion: v1
kind: Service
metadata:
  name: zk-hs
  labels:
    app: zk
  namespace: kafka
spec:
  selector:
    app: zk
  clusterIP: None
  ports:
    - name: server
      port: 2888
    - name: leader-election
      port: 3888
---
apiVersion: v1
kind: Service
metadata:
  name: zk-cs
  labels:
    app: zk
  namespace: kafka
spec:
  selector:
    app: zk
  type: NodePort
  ports:
    - name: client
      port: 2181
      nodePort: 31811
---
apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
  name: zk-pdb
  namespace: kafka
spec:
  selector:
    matchLabels:
      app: zk
  maxUnavailable: 1
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: zk
  namespace: kafka
spec:
  serviceName: "zk-hs"
  replicas: 3 # by default is 1
  selector:
    matchLabels:
      app: zk # has to match .spec.template.metadata.labels
  updateStrategy:
    type: RollingUpdate
  podManagementPolicy: Parallel
  template:
    metadata:
      labels:
        app: zk # has to match .spec.selector.matchLabels
    spec:
      containers:
        - name: zk
          imagePullPolicy: Always
          image: leolee32/kubernetes-library:kubernetes-zookeeper1.0-3.4.10
          resources:
            requests:
              memory: "2G"
              cpu: "1"
          ports:
            - containerPort: 2181
              name: client
            - containerPort: 2888
              name: server
            - containerPort: 3888
              name: leader-election
          command:
            - sh
            - -c
            - "start-zookeeper \
        --servers=3 \
        --data_dir=/var/lib/zookeeper/data \
        --data_log_dir=/var/lib/zookeeper/data/log \
        --conf_dir=/opt/zookeeper/conf \
        --client_port=2181 \
        --election_port=3888 \
        --server_port=2888 \
        --tick_time=2000 \
        --init_limit=10 \
        --sync_limit=5 \
        --heap=4G \
        --max_client_cnxns=60 \
        --snap_retain_count=3 \
        --purge_interval=12 \
        --max_session_timeout=40000 \
        --min_session_timeout=4000 \
        --log_level=INFO"
          readinessProbe:
            exec:
              command:
                - sh
                - -c
                - "zookeeper-ready 2181"
            initialDelaySeconds: 10
            timeoutSeconds: 5
          livenessProbe:
            exec:
              command:
                - sh
                - -c
                - "zookeeper-ready 2181"
            initialDelaySeconds: 10
            timeoutSeconds: 5
          volumeMounts:
            - name: datadir
              mountPath: /var/lib/zookeeper
  volumeClaimTemplates:
    - metadata:
        name: datadir
        annotations:
          volume.beta.kubernetes.io/storage-class: managed-nfs-storage
      spec:
        accessModes: [ "ReadWriteOnce" ]
        resources:
          requests:
            storage: 1Gi

安装kafka

采用helm的方式安装kafka最简单,真的不要太方便了,只需要简单的配置即可

借鉴以下文章的方式,镜像采用的是bitnami/kafka
https://artifacthub.io/packages/helm/bitnami/kafka
https://www.shuzhiduo.com/A/n2d9qvb0dD/

下载kafka的chart：bitnami/kafka
下载成功后当前目录有个压缩包：kafka-0.20.8.tgz，解压：tar -zxvf kafka-0.20.8.tgz
进入解压后的kafka目录，编辑values.yaml文件，下面是具体的修改点：

校验安装是否成功

1. kubectl exec -ti zk-1 -n kafka -- zkCli.sh -server zk-cs:2181

出现
Welcome to ZooKeeper!
JLine support is enabled
即安装成功

2. kubectl exec -ti kafka-0 -n kafka -- /bin/bash

登录到pod里面
执行获取topic : kafka-topics.sh --list --zookeeper zk-cs:2181
kafka-topics.sh --list --zookeeper 10.45.2.13:31811
创建topic : kafka-topics.sh --create --topic testtopic --zookeeper zk-cs:2181 --partitions 1 --replication-factor 1
执行获取topic : kafka-topics.sh --list --zookeeper zk-cs:2181
往topic中生产数据: kafka-console-producer.sh --broker-list nodeport:31092 --topic wisecotest

kafka-2.kafka-cs.kafka.svc.cluster.local:31092
// 对外暴露的地址端口10.233.137.116:9092 32092
往topic中生产数据: kafka-console-producer.sh --broker-list kafka-cs:9092 --topic wisecotest

从topic中获取数据: kafka-console-consumer.sh --bootstrap-server kafka-2.kafka-cs.kafka.svc.cluster.local:31092 --topic wisecotest --from-beginning

安装问题解答

1. 出现: 0/3 nodes are available: 1 node(s) had taint {node-role.kubernetes.io/master: }, that the pod didn't tolerate, 2 node(s) didn't match pod anti-affinity rules.

原因是设定了亲和性,导致每个节点上只能安装一个类型,master节点又不允许安装其他的节点,导致失败

2. 注意事项,采用stateful set的方式安装kafka在集群内部是可以访问的,但是集群外是不能访问

stateful set 安装kafka后
通过service对外暴露的ip:port能访问到zookeeper,zookeeper返回的是注册的集群内部的地址(这里应该可以修改),导致当前链接不上返回的地址
最后采用了helm的安装方式,等于是对外暴露了三个service,每个service对应pod
然后在zookeeper中看到的注册地址中有个external的存在