「后浪云」Apache Livy on k8s 讲解与实战操作

一、概述

Livy是一个提供Rest接口和spark集群交互的服务。它可以提交Spark Job或者Spark一段代码,同步或者异步的返回结果;也提供Sparkcontext的管理,通过Restful接口或RPC客户端库。Livy也简化了与Spark与应用服务的交互,这允许通过web/mobile与Spark的使用交互。

官网：https://livy.incubator.apache.org/GitHub地址：https://github.com/apache/incubator-livy关于Apache Livy更多介绍也可以参考我这篇文章：Spark开源REST服务——Apache Livy（Spark 客户端）

二、开始编排部署

1.部署包准备

这里也提供上面编译好的livy部署包，有需要的小伙伴可以自行下载：

链接：https://pan.baidu.com/s/1pPCbe0lUJ6ji8rvQYsVw9A?pwd=qn7i提取码：qn7i

1）构建镜像

Dockerfile

FROM myharbor.com/bigdata/centos:7.9.2009

RUN rm -f /etc/localtime && ln -sv /usr/share/zoneinfo/Asia/Shanghai /etc/localtime && echo "Asia/Shanghai">/etc/timezone
RUN export LANG=zh_CN.UTF-8

### install tools
RUN yum install -y vim tar wget curl less telnet net-tools lsof

RUN groupadd --system --gid=9999 admin && useradd --system -m /home/admin --uid=9999 --gid=admin admin

RUN mkdir -p /opt/apache

ADD apache-livy-0.8.0-incubating-SNAPSHOT-bin.zip/opt/apache/
ENV LIVY_HOME=/opt/apache/apache-livy
RUN ln -s /opt/apache/apache-livy-0.8.0-incubating-SNAPSHOT-bin $LIVY_HOME

ADD hadoop-3.3.2.tar.gz/opt/apache/
ENV HADOOP_HOME=/opt/apache/hadoop
RUN ln -s /opt/apache/hadoop-3.3.2 $HADOOP_HOME
ENV HADOOP_CONFIG_DIR=${HADOOP_HOME}/etc/hadoop

ADD spark-3.3.0-bin-hadoop3.tar.gz/opt/apache/
ENV SPARK_HOME=/opt/apache/spark
RUN ln -s /opt/apache/spark-3.3.0-bin-hadoop3 $SPARK_HOME

ENV PATH=${LIVY_HOME}/bin:${HADOOP_HOME}/bin:${SPARK_HOME}/bin:$PATH

RUN chown -R admin:admin /opt/apache

WORKDIR $LIVY_HOME

ENTRYPOINT ${LIVY_HOME}/bin/livy-server start;tail -f ${LIVY_HOME}/logs/livy-root-server.out

【注意】hadoop包里的core-site.xml，hdfs-site.xml，yarn-site.xml

开始构建镜像

docker build -t myharbor.com/bigdata/livy:0.8.0 . --no-cache

### 参数解释
# -t：指定镜像名称
# . ：当前目录Dockerfile
# -f：指定Dockerfile路径
#  --no-cache：不缓存

# 推送到harbor
docker push myharbor.com/bigdata/livy:0.8.0

2）创建livy chart模板

helm create livy

3）修改yaml编排

livy/values.yaml

replicaCount:1

image:
  repository: myharbor.com/bigdata/livy
  pullPolicy: IfNotPresent
  # Overrides the image tag whose default is the chart appVersion.
  tag:"0.8.0"

securityContext:
  runAsUser:9999
  runAsGroup:9999
  privileged:true

service:
  type: NodePort
  port:8998
  nodePort:31998

livy/templates/configmap.yaml

apiVersion: v1
kind: ConfigMap
metadata:
  name:{{ include "livy.fullname" . }}
  labels:
{{- include "livy.labels" . | nindent 4}}
data:
  livy.conf:|-
    livy.spark.master= yarn
    livy.spark.deploy-mode = client
    livy.environment= production
    livy.impersonation.enabled=true
    livy.server.csrf_protection.enabled=false
    livy.server.port={{.Values.service.port}}
    livy.server.session.timeout=3600000
    livy.server.recovery.mode= recovery
    livy.server.recovery.state-store = filesystem
    livy.server.recovery.state-store.url=/tmp/livy
    livy.repl.enable-hive-context =true
  livy-env.sh:|-
    export JAVA_HOME=/opt/apache/jdk1.8.0_212
    export HADOOP_HOME=/opt/apache/hadoop
    export HADOOP_CONF_DIR=/opt/apache/hadoop/etc/hadoop
    export SPARK_HOME=/opt/apache/spark
    export SPARK_CONF_DIR=/opt/apache/spark/conf
    export LIVY_LOG_DIR=/opt/apache/livy/logs
    export LIVY_PID_DIR=/opt/apache/livy/pid-dir
    export LIVY_SERVER_JAVA_OPTS="-Xmx512m"
  spark-blacklist.conf:|-
    spark.master
    spark.submit.deployMode

    # Disallow overriding the location of Spark cached jars.
    spark.yarn.jar
    spark.yarn.jars
    spark.yarn.archive

    # Don't allow users to override the RSC timeout.
    livy.rsc.server.idle-timeout

livy/templates/deployment.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name:{{ include "livy.fullname" . }}
  labels:
{{- include "livy.labels" . | nindent 4}}
spec:
{{- if not.Values.autoscaling.enabled}}
  replicas:{{.Values.replicaCount}}
{{- end }}
  selector:
    matchLabels:
{{- include "livy.selectorLabels" . | nindent 6}}
  template:
    metadata:
{{- with .Values.podAnnotations}}
      annotations:
{{- toYaml . | nindent 8}}
{{- end }}
      labels:
{{- include "livy.selectorLabels" . | nindent 8}}
    spec:
{{- with .Values.imagePullSecrets}}
      imagePullSecrets:
{{- toYaml . | nindent 8}}
{{- end }}
      serviceAccountName:{{ include "livy.serviceAccountName" . }}
      securityContext:
{{- toYaml .Values.podSecurityContext| nindent 8}}
      containers:
- name:{{.Chart.Name}}
          securityContext:
{{- toYaml .Values.securityContext| nindent 12}}
          image:"{{ .Values.image.repository }}:{{ .Values.image.tag | default .Chart.AppVersion }}"
          imagePullPolicy:{{.Values.image.pullPolicy}}
          ports:
- name: http
              containerPort:8998
              protocol: TCP
          livenessProbe:
            httpGet:
              path:/
              port: http
          readinessProbe:
            httpGet:
              path:/
              port: http
          resources:
{{- toYaml .Values.resources| nindent 12}}
{{- with .Values.securityContext}}
          securityContext:
            runAsUser:{{.runAsUser}}
            runAsGroup:{{.runAsGroup}}
            privileged:{{.privileged}}
{{- end }}
          volumeMounts:
- name:{{.Release.Name}}-livy-conf
              mountPath:/opt/apache/livy/conf/livy.conf
              subPath: livy.conf
- name:{{.Release.Name}}-livy-env
              mountPath:/opt/apache/livy/conf/livy-env.sh
              subPath: livy-env.sh
- name:{{.Release.Name}}-spark-blacklist-conf
              mountPath:/opt/apache/livy/conf/spark-blacklist.conf
              subPath: spark-blacklist.conf
{{- with .Values.nodeSelector}}
      nodeSelector:
{{- toYaml . | nindent 8}}
{{- end }}
{{- with .Values.affinity}}
      affinity:
{{- toYaml . | nindent 8}}
{{- end }}
{{- with .Values.tolerations}}
      tolerations:
{{- toYaml . | nindent 8}}
{{- end }}
      volumes:
- name:{{.Release.Name}}-livy-conf
          configMap:
            name:{{ include "livy.fullname" . }}
- name:{{.Release.Name}}-livy-env
          configMap:
            name:{{ include "livy.fullname" . }}
- name:{{.Release.Name}}-spark-blacklist-conf
          configMap:
            name:{{ include "livy.fullname" . }}

4）开始部署

helm install livy ./livy -n livy --create-namespace

NOTES

NOTES:
1. Get the application URL by running these commands:
  export NODE_PORT=$(kubectl get --namespace livy -o jsonpath="{.spec.ports[0].nodePort}" services livy)
  export NODE_IP=$(kubectl get nodes --namespace livy -o jsonpath="{.items[0].status.addresses[0].address}")
  echo http://$NODE_IP:$NODE_PORT

查看

kubectl get pods,svc -n livy -owide

web地址：http://192.168.182.110:31998/ui

5）测试验证

curl -s -XPOST -d '{"file":"hdfs://myhdfs/tmp/spark-examples_2.12-3.3.0.jar","className":"org.apache.spark.examples.SparkPi","name":"SparkPi-test"}'-H "Content-Type: application/json"  http://local-168-182-110:31998/batches|python -m json.tool

6）卸载

helm uninstall livy -n livy

git地址：https://gitee.com/hadoop-bigdata/livy-on-k8s​