• 14 张图详解 Zookeeper + Kafka on K8S 环境部署


    转载自:https://www.cnblogs.com/liugp/p/16706192.html

    一、概述

    • Apache ZooKeeper 是一个集中式服务,用于维护配置信息、命名、提供分布式同步和提供组服务,ZooKeeper 致力于开发和维护一个开源服务器,以实现高度可靠的分布式协调,其实也可以认为就是一个分布式数据库,只是结构比较特殊,是树状结构。官网文档:

      https://zookeeper.apache.org/doc/r3.8.0/

    图片

    • Kafka是最初由 Linkedin 公司开发,是一个分布式、支持分区的(partition)、多副本的(replica),基于 zookeeper 协调的分布式消息系统。官方文档:

      https://kafka.apache.org/documentation/

    图片

    二、Zookeeper on k8s 部署

    图片

    1)添加源

    部署包地址:

    https://artifacthub.io/packages/helm/zookeeper/zookeeper

    helm repo add bitnami https://charts.bitnami.com/bitnami
    helm pull bitnami/zookeeper
    tar -xf  zookeeper-10.2.1.tgz

    2)修改配置

    • 修改zookeeper/values.yaml
    image:
      registry: myharbor.com
      repository: bigdata/zookeeper
      tag: 3.8.0-debian-11-r36
    ...

    replicaCount: 3

    ...

    service:
      type: NodePort
      nodePorts:
        #NodePort 默认范围是 30000-32767
        client: "32181"
        tls: "32182"

    ...

    persistence:
      storageClass: "zookeeper-local-storage"
      size: "10Gi"
      # 目录需要提前在宿主机上创建
      local:
        - name: zookeeper-0
          host: "local-168-182-110"
          path: "/opt/bigdata/servers/zookeeper/data/data1"
        - name: zookeeper-1
          host: "local-168-182-111"
          path: "/opt/bigdata/servers/zookeeper/data/data1"
        - name: zookeeper-2
          host: "local-168-182-112"
          path: "/opt/bigdata/servers/zookeeper/data/data1"

    ...

    # Enable Prometheus to access ZooKeeper metrics endpoint
    metrics:
      enabled: true
    • 添加zookeeper/templates/pv.yaml
    {{- range .Values.persistence.local }}
    ---
    apiVersion: v1
    kind: PersistentVolume
    metadata:
      name: {{ .name }}
      labels:
        name: {{ .name }}
    spec:
      storageClassName: {{ $.Values.persistence.storageClass }}
      capacity:
        storage: {{ $.Values.persistence.size }}
      accessModes:
        - ReadWriteOnce
      local:
        path: {{ .path }}
      nodeAffinity:
        required:
          nodeSelectorTerms:
            - matchExpressions:
                - key: kubernetes.io/hostname
                  operator: In
                  values:
                    - {{ .host }}
    ---
    {{- end }}

    • 添加zookeeper/templates/storage-class.yaml
    kind: StorageClass
    apiVersion: storage.k8s.io/v1
    metadata:
      name: {{ .Values.persistence.storageClass }}
    provisioner: kubernetes.io/no-provisioner

    3)开始安装

    # 先准备好镜像
    docker pull docker.io/bitnami/zookeeper:3.8.0-debian-11-r36
    docker tag docker.io/bitnami/zookeeper:3.8.0-debian-11-r36 myharbor.com/bigdata/zookeeper:3.8.0-debian-11-r36
    docker push myharbor.com/bigdata/zookeeper:3.8.0-debian-11-r36

    # 开始安装
    helm install zookeeper ./zookeeper -n zookeeper --create-namespace

    NOTES

    NAME: zookeeper
    LAST DEPLOYED: Sun Sep 18 18:24:03 2022
    NAMESPACE: zookeeper
    STATUS: deployed
    REVISION: 1
    TEST SUITE: None
    NOTES:
    CHART NAME: zookeeper
    CHART VERSION: 10.2.1
    APP VERSION: 3.8.0

    ** Please be patient while the chart is being deployed **

    ZooKeeper can be accessed via port 2181 on the following DNS name from within your cluster:

        zookeeper.zookeeper.svc.cluster.local

    To connect to your ZooKeeper server run the following commands:

        export POD_NAME=$(kubectl get pods --namespace zookeeper -l "app.kubernetes.io/name=zookeeper,app.kubernetes.io/instance=zookeeper,app.kubernetes.io/component=zookeeper" -o jsonpath="{.items[0].metadata.name}")
        kubectl exec -it $POD_NAME -- zkCli.sh

    To connect to your ZooKeeper server from outside the cluster execute the following commands:

        export NODE_IP=$(kubectl get nodes --namespace zookeeper -o jsonpath="{.items[0].status.addresses[0].address}")
        export NODE_PORT=$(kubectl get --namespace zookeeper -o jsonpath="{.spec.ports[0].nodePort}" services zookeeper)
        zkCli.sh $NODE_IP:$NODE_PORT

    图片
    查看 pod 状态

    kubectl get pods,svc -n zookeeper -owide

    图片

    4)测试验证

    # 登录zookeeper pod
    kubectl exec -it zookeeper-0 -n zookeeper -- zkServer.sh status
    kubectl exec -it zookeeper-1 -n zookeeper -- zkServer.sh status
    kubectl exec -it zookeeper-2 -n zookeeper -- zkServer.sh status

    kubectl exec -it zookeeper-0 -n zookeeper -- bash

    图片

    5)Prometheus 监控

    Prometheus:https://prometheus.k8s.local/targets?search=zookeeper
    图片

    可以通过命令查看采集数据

    kubectl get --raw http://10.244.0.52:9141/metrics
    kubectl get --raw http://10.244.1.101:9141/metrics
    kubectl get --raw http://10.244.2.137:9141/metrics

    Grafana:https://grafana.k8s.local/
    账号:admin,密码通过下面命令获取

    kubectl get secret --namespace grafana grafana -o jsonpath="{.data.admin-password}" | base64 --decode ; echo

    导入 grafana 模板,集群资源监控:10465
    官方模块下载地址:

    https://grafana.com/grafana/dashboards/

    图片

    6)卸载

    helm uninstall zookeeper -n zookeeper

    kubectl delete pod -n zookeeper `kubectl get pod -n zookeeper|awk 'NR>1{print $1}'` --force
    kubectl patch ns zookeeper -p '{"metadata":{"finalizers":null}}'
    kubectl delete ns zookeeper --force

    三、Kafka on k8s 部署

    1)添加源

    部署包地址:

    https://artifacthub.io/packages/helm/bitnami/kafka

    helm repo add bitnami https://charts.bitnami.com/bitnami
    helm pull bitnami/kafka
    tar -xf kafka-18.4.2.tgz

    2)修改配置

    • 修改kafka/values.yaml
    image:
      registry: myharbor.com
      repository: bigdata/kafka
      tag: 3.2.1-debian-11-r16

    ...

    replicaCount: 3

    ...

    service:
      type: NodePort
      nodePorts:
        client: "30092"
        external: "30094"

    ...

    externalAccess
      enabled: true
      service:
        type: NodePort
         nodePorts:
           - 30001
           - 30002
           - 30003
         useHostIPs: true

    ...

    persistence:
      storageClass: "kafka-local-storage"
      size: "10Gi"
      # 目录需要提前在宿主机上创建
      local:
        - name: kafka-0
          host: "local-168-182-110"
          path: "/opt/bigdata/servers/kafka/data/data1"
        - name: kafka-1
          host: "local-168-182-111"
          path: "/opt/bigdata/servers/kafka/data/data1"
        - name: kafka-2
          host: "local-168-182-112"
          path: "/opt/bigdata/servers/kafka/data/data1"

    ...

    metrics:
      kafka:
        enabled: true
        image:
          registry: myharbor.com
          repository: bigdata/kafka-exporter
          tag: 1.6.0-debian-11-r8
        jmx:
          enabled: true
          image:
          registry: myharbor.com
          repository: bigdata/jmx-exporter
          tag: 0.17.1-debian-11-r1
          annotations:
            prometheus.io/path: "/metrics"

    ...

    zookeeper:
      enabled: false

    ...

    externalZookeeper
      servers:
        - zookeeper-0.zookeeper-headless.zookeeper
        - zookeeper-1.zookeeper-headless.zookeeper
        - zookeeper-2.zookeeper-headless.zookeeper
    • 添加kafka/templates/pv.yaml
    {{- range .Values.persistence.local }}
    ---
    apiVersion: v1
    kind: PersistentVolume
    metadata:
      name: {{ .name }}
      labels:
        name: {{ .name }}
    spec:
      storageClassName: {{ $.Values.persistence.storageClass }}
      capacity:
        storage: {{ $.Values.persistence.size }}
      accessModes:
        - ReadWriteOnce
      local:
        path: {{ .path }}
      nodeAffinity:
        required:
          nodeSelectorTerms:
            - matchExpressions:
                - key: kubernetes.io/hostname
                  operator: In
                  values:
                    - {{ .host }}
    ---
    {{- end }}
    • 添加kafka/templates/storage-class.yaml
    kind: StorageClass
    apiVersion: storage.k8s.io/v1
    metadata:
      name: {{ .Values.persistence.storageClass }}
    provisioner: kubernetes.io/no-provisioner

    3)开始安装

    # 先准备好镜像
    docker pull docker.io/bitnami/kafka:3.2.1-debian-11-r16
    docker tag docker.io/bitnami/kafka:3.2.1-debian-11-r16 myharbor.com/bigdata/kafka:3.2.1-debian-11-r16
    docker push myharbor.com/bigdata/kafka:3.2.1-debian-11-r16

    # node-export
    docker pull docker.io/bitnami/kafka-exporter:1.6.0-debian-11-r8
    docker tag docker.io/bitnami/kafka-exporter:1.6.0-debian-11-r8 myharbor.com/bigdata/kafka-exporter:1.6.0-debian-11-r8
    docker push myharbor.com/bigdata/kafka-exporter:1.6.0-debian-11-r8

    # JXM
    docker.io/bitnami/jmx-exporter:0.17.1-debian-11-r1
    docker tag docker.io/bitnami/jmx-exporter:0.17.1-debian-11-r1 myharbor.com/bigdata/jmx-exporter:0.17.1-debian-11-r1
    docker push myharbor.com/bigdata/jmx-exporter:0.17.1-debian-11-r1

    #开始安装
    helm install kafka ./kafka -n kafka --create-namespace

    NOTES

    NAME: kafka
    LAST DEPLOYED: Sun Sep 18 20:57:02 2022
    NAMESPACE: kafka
    STATUS: deployed
    REVISION: 1
    TEST SUITE: None
    NOTES:
    CHART NAME: kafka
    CHART VERSION: 18.4.2
    APP VERSION: 3.2.1
    ---------------------------------------------------------------------------------------------
     WARNING

        By specifying "serviceType=LoadBalancer" and not configuring the authentication
        you have most likely exposed the Kafka service externally without any
        authentication mechanism.

        For security reasons, we strongly suggest that you switch to "ClusterIP" or
        "NodePort". As alternative, you can also configure the Kafka authentication.

    ---------------------------------------------------------------------------------------------

    ** Please be patient while the chart is being deployed **

    Kafka can be accessed by consumers via port 9092 on the following DNS name from within your cluster:

        kafka.kafka.svc.cluster.local

    Each Kafka broker can be accessed by producers via port 9092 on the following DNS name(s) from within your cluster:

        kafka-0.kafka-headless.kafka.svc.cluster.local:9092
        kafka-1.kafka-headless.kafka.svc.cluster.local:9092
        kafka-2.kafka-headless.kafka.svc.cluster.local:9092

    To create a pod that you can use as a Kafka client run the following commands:

        kubectl run kafka-client --restart='Never' --image docker.io/bitnami/kafka:3.2.1-debian-11-r16 --namespace kafka --command -- sleep infinity
        kubectl exec --tty -i kafka-client --namespace kafka -- bash

        PRODUCER:
            kafka-console-producer.sh \

                --broker-list kafka-0.kafka-headless.kafka.svc.cluster.local:9092,kafka-1.kafka-headless.kafka.svc.cluster.local:9092,kafka-2.kafka-headless.kafka.svc.cluster.local:9092 \
                --topic test

        CONSUMER:
            kafka-console-consumer.sh \

                --bootstrap-server kafka.kafka.svc.cluster.local:9092 \
                --topic test \
                --from-beginning

    To connect to your Kafka server from outside the cluster, follow the instructions below:

        Kafka brokers domain: You can get the external node IP from the Kafka configuration file with the following commands (Check the EXTERNAL listener)

            1. Obtain the pod name:

            kubectl get pods --namespace kafka -l "app.kubernetes.io/name=kafka,app.kubernetes.io/instance=kafka,app.kubernetes.io/component=kafka"

            2. Obtain pod configuration:

            kubectl exec -it KAFKA_POD -- cat /opt/bitnami/kafka/config/server.properties | grep advertised.listeners

        Kafka brokers port: You will have a different node port for each Kafka broker. You can get the list of configured node ports using the command below:

            echo "$(kubectl get svc --namespace kafka -l "app.kubernetes.io/name=kafka,app.kubernetes.io/instance=kafka,app.kubernetes.io/component=kafka,pod" -o jsonpath='{.items[*].spec.ports[0].nodePort}' | tr ' ' '\n')"

    图片
    查看 pod 状态

    kubectl get pods,svc -n kafka -owide

    图片

    4)测试验证

    # 登录zookeeper pod
    kubectl exec -it kafka-0 -n kafka -- bash

    1、创建 Topic(一个副本一个分区)

    --create: 指定创建topic动作

    --topic:指定新建topic的名称

    --bootstrap-server: 指定kafka连接地址

    --config:指定当前topic上有效的参数值,参数列表参考文档为: Topic-level configuration

    --partitions:指定当前创建的kafka分区数量,默认为1个

    --replication-factor:指定每个分区的复制因子个数,默认1个
    kafka-topics.sh --create --topic test001 --bootstrap-server kafka.kafka:9092 --partitions 1 --replication-factor 1
    # 查看
    kafka-topics.sh --describe --bootstrap-server kafka.kafka:9092  --topic test001

    图片

    2、查看 Topic 列表

    kafka-topics.sh --list --bootstrap-server kafka.kafka:9092

    3、生产者/消费者测试

    【生产者】

    kafka-console-producer.sh --broker-list kafka.kafka:9092 --topic test001

    {"id":"1","name":"n1","age":"20"}
    {"id":"2","name":"n2","age":"21"}
    {"id":"3","name":"n3","age":"22"}

    【消费者】

    # 从头开始消费
    kafka-console-consumer.sh --bootstrap-server kafka.kafka:9092 --topic test001 --from-beginning
    # 指定从分区的某个位置开始消费,这里只指定了一个分区,可以多写几行或者遍历对应的所有分区
    kafka-console-consumer.sh --bootstrap-server kafka.kafka:9092 --topic test001 --partition 0 --offset 100 --group test001

    4、查看数据积压

    kafka-consumer-groups.sh --bootstrap-server kafka.kafka:9092 --describe --group test001

    5、删除 topic

    kafka-topics.sh --delete --topic test001 --bootstrap-server kafka.kafka:9092

    5)Prometheus 监控

    Prometheus:

    https://prometheus.k8s.local/targets?search=kafka

    图片

    可以通过命令查看采集数据

    kubectl get --raw http://10.244.2.165:9308/metrics

    Grafana:https://grafana.k8s.local/
    账号:admin,密码通过下面命令获取

    kubectl get secret --namespace grafana grafana -o jsonpath="{.data.admin-password}" | base64 --decode ; echo

    导入 grafana 模板,集群资源监控:11962
    官方模块下载地址:

    https://grafana.com/grafana/dashboards/

    6)卸载

    helm uninstall kafka -n kafka

    kubectl delete pod -n kafka `kubectl get pod -n kafka|awk 'NR>1{print $1}'` --force
    kubectl patch ns kafka  -p '{"metadata":{"finalizers":null}}'
    kubectl delete ns kafka  --force

    zookeeper + kafka on k8s 环境部署 就先到这里了,小伙伴有任何疑问,欢迎给我留言!

    链接:https://www.cnblogs.com/liugp/p/16706192.html

  • 相关阅读:
    Activity详解
    Log和LogCat的使用
    Android Studio项目目录结构
    Android系统架构
    [SCOI2016]美味
    [SCOI2016]背单词
    [SCOI2016]幸运数字
    [BZOJ4170]极光
    [JSOI2016]扭动的回文串
    [SCOI2016]萌萌哒
  • 原文地址:https://www.cnblogs.com/testzcy/p/16709915.html
Copyright © 2020-2023  润新知