将 Kafka 主题流式传输到 Hive


Apache Kafka 是一个开源分布式流式传输平台,用于实时数据流水线和数据集成。它提供了一个高效且可扩缩的流式传输系统,可用于各种应用,包括:

  • 实时分析
  • 流处理
  • 日志汇总
  • 分布式消息传递
  • 事件流处理

目标

  1. 在带有 ZooKeeper 的 Dataproc 高可用性集群(在本教程中称为“Dataproc Kafka 集群”)上安装 Kafka。

  2. 创建虚构的客户数据,然后将该数据发布到 Kafka 主题。

  3. 在 Cloud Storage 中创建 Hive Parquet 和 ORC 表,以接收流式传输的 Kafka 主题数据。

  4. 提交 PySpark 作业,以订阅 Kafka 主题并以 Parquet 和 ORC 格式将该主题流式传输到 Cloud Storage。

  5. 对流式传输的 Hive 表数据运行查询,以统计流式传输的 Kafka 消息数量。

费用

在本文档中,您将使用 Google Cloud的以下收费组件:

如需根据您的预计使用量来估算费用,请使用价格计算器

新 Google Cloud 用户可能有资格申请免费试用

完成本文档中描述的任务后,您可以通过删除所创建的资源来避免继续计费。如需了解详情,请参阅清理

准备工作

如果尚未创建 Google Cloud 项目,请创建一个。

  1. Sign in to your Google Cloud account. If you're new to Google Cloud, create an account to evaluate how our products perform in real-world scenarios. New customers also get $300 in free credits to run, test, and deploy workloads.
  2. In the Google Cloud console, on the project selector page, select or create a Google Cloud project.

    Roles required to select or create a project

    • Select a project: Selecting a project doesn't require a specific IAM role—you can select any project that you've been granted a role on.
    • Create a project: To create a project, you need the Project Creator (roles/resourcemanager.projectCreator), which contains the resourcemanager.projects.create permission. Learn how to grant roles.

    Go to project selector

  3. Verify that billing is enabled for your Google Cloud project.

  4. Enable the Dataproc, Compute Engine, and Cloud Storage APIs.

    Roles required to enable APIs

    To enable APIs, you need the Service Usage Admin IAM role (roles/serviceusage.serviceUsageAdmin), which contains the serviceusage.services.enable permission. Learn how to grant roles.

    Enable the APIs

  5. In the Google Cloud console, on the project selector page, select or create a Google Cloud project.

    Roles required to select or create a project

    • Select a project: Selecting a project doesn't require a specific IAM role—you can select any project that you've been granted a role on.
    • Create a project: To create a project, you need the Project Creator (roles/resourcemanager.projectCreator), which contains the resourcemanager.projects.create permission. Learn how to grant roles.

    Go to project selector

  6. Verify that billing is enabled for your Google Cloud project.

  7. Enable the Dataproc, Compute Engine, and Cloud Storage APIs.

    Roles required to enable APIs

    To enable APIs, you need the Service Usage Admin IAM role (roles/serviceusage.serviceUsageAdmin), which contains the serviceusage.services.enable permission. Learn how to grant roles.

    Enable the APIs

  8. In the Google Cloud console, go to the Cloud Storage Buckets page.

    Go to Buckets

  9. Click Create.
  10. On the Create a bucket page, enter your bucket information. To go to the next step, click Continue.
    1. In the Get started section, do the following:
      • Enter a globally unique name that meets the bucket naming requirements.
      • To add a bucket label, expand the Labels section (), click Add label, and specify a key and a value for your label.
    2. In the Choose where to store your data section, do the following:
      1. Select a Location type.
      2. Choose a location where your bucket's data is permanently stored from the Location type drop-down menu.
      3. To set up cross-bucket replication, select Add cross-bucket replication via Storage Transfer Service and follow these steps:

        Set up cross-bucket replication

        1. In the Bucket menu, select a bucket.
        2. In the Replication settings section, click Configure to configure settings for the replication job.

          The Configure cross-bucket replication pane appears.

          • To filter objects to replicate by object name prefix, enter a prefix that you want to include or exclude objects from, then click Add a prefix.
          • To set a storage class for the replicated objects, select a storage class from the Storage class menu. If you skip this step, the replicated objects will use the destination bucket's storage class by default.
          • Click Done.
    3. In the Choose how to store your data section, do the following:
      1. Select a default storage class for the bucket or Autoclass for automatic storage class management of your bucket's data.
      2. To enable hierarchical namespace, in the Optimize storage for data-intensive workloads section, select Enable hierarchical namespace on this bucket.
    4. In the Choose how to control access to objects section, select whether or not your bucket enforces public access prevention, and select an access control method for your bucket's objects.
    5. In the Choose how to protect object data section, do the following:
      • Select any of the options under Data protection that you want to set for your bucket.
        • To enable soft delete, click the Soft delete policy (For data recovery) checkbox, and specify the number of days you want to retain objects after deletion.
        • To set Object Versioning, click the Object versioning (For version control) checkbox, and specify the maximum number of versions per object and the number of days after which the noncurrent versions expire.
        • To enable the retention policy on objects and buckets, click the Retention (For compliance) checkbox, and then do the following:
          • To enable Object Retention Lock, click the Enable object retention checkbox.
          • To enable Bucket Lock, click the Set bucket retention policy checkbox, and choose a unit of time and a length of time for your retention period.
      • To choose how your object data will be encrypted, expand the Data encryption section (), and select a Data encryption method.
  11. Click Create.
  12. 教程步骤

    请执行以下步骤创建 Dataproc Kafka 集群,以 Parquet 或 ORC 格式将 Kafka 主题读入 Cloud Storage。

    将 Kafka 安装脚本复制到 Cloud Storage

    kafka.sh 初始化操作脚本会在 Dataproc 集群上安装 Kafka。

    1. 浏览代码。

      #!/bin/bash
      #    Copyright 2015 Google, Inc.
      #
      #    Licensed under the Apache License, Version 2.0 (the "License");
      #    you may not use this file except in compliance with the License.
      #    You may obtain a copy of the License at
      #
      #        http://www.apache.org/licenses/LICENSE-2.0
      #
      #    Unless required by applicable law or agreed to in writing, software
      #    distributed under the License is distributed on an "AS IS" BASIS,
      #    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
      #    See the License for the specific language governing permissions and
      #    limitations under the License.
      #
      # This script installs Apache Kafka (http://kafka.apache.org) on a Google Cloud
      # Dataproc cluster.
      
      set -euxo pipefail
      
      readonly ZOOKEEPER_HOME=/usr/lib/zookeeper
      readonly KAFKA_HOME=/usr/lib/kafka
      readonly KAFKA_PROP_FILE='/etc/kafka/conf/server.properties'
      readonly ROLE="$(/usr/share/google/get_metadata_value attributes/dataproc-role)"
      readonly RUN_ON_MASTER="$(/usr/share/google/get_metadata_value attributes/run-on-master || echo false)"
      readonly KAFKA_ENABLE_JMX="$(/usr/share/google/get_metadata_value attributes/kafka-enable-jmx || echo false)"
      readonly KAFKA_JMX_PORT="$(/usr/share/google/get_metadata_value attributes/kafka-jmx-port || echo 9999)"
      readonly INSTALL_KAFKA_PYTHON="$(/usr/share/google/get_metadata_value attributes/install-kafka-python || echo false)"
      
      # The first ZooKeeper server address, e.g., "cluster1-m-0:2181".
      ZOOKEEPER_ADDRESS=''
      # Integer broker ID of this node, e.g., 0
      BROKER_ID=''
      
      function retry_apt_command() {
        cmd="$1"
        for ((i = 0; i < 10; i++)); do
          if eval "$cmd"; then
            return 0
          fi
          sleep 5
        done
        return 1
      }
      
      function recv_keys() {
        retry_apt_command "apt-get install -y gnupg2 &&\
                           apt-key adv --keyserver keyserver.ubuntu.com --recv-keys B7B3B788A8D3785C"
      }
      
      function update_apt_get() {
        retry_apt_command "apt-get update"
      }
      
      function install_apt_get() {
        pkgs="$@"
        retry_apt_command "apt-get install -y $pkgs"
      }
      
      function err() {
        echo "[$(date +'%Y-%m-%dT%H:%M:%S%z')]: $@" >&2
        return 1
      }
      
      # Returns the list of broker IDs registered in ZooKeeper, e.g., " 0, 2, 1,".
      function get_broker_list() {
        ${KAFKA_HOME}/bin/zookeeper-shell.sh "${ZOOKEEPER_ADDRESS}" \
          <<<"ls /brokers/ids" |
          grep '\[.*\]' |
          sed 's/\[/ /' |
          sed 's/\]/,/'
      }
      
      # Waits for zookeeper to be up or time out.
      function wait_for_zookeeper() {
        for i in {1..20}; do
          if "${ZOOKEEPER_HOME}/bin/zkCli.sh" -server "${ZOOKEEPER_ADDRESS}" ls /; then
            return 0
          else
            echo "Failed to connect to ZooKeeper ${ZOOKEEPER_ADDRESS}, retry ${i}..."
            sleep 5
          fi
        done
        echo "Failed to connect to ZooKeeper ${ZOOKEEPER_ADDRESS}" >&2
        exit 1
      }
      
      # Wait until the current broker is registered or time out.
      function wait_for_kafka() {
        for i in {1..20}; do
          local broker_list=$(get_broker_list || true)
          if [[ "${broker_list}" == *" ${BROKER_ID},"* ]]; then
            return 0
          else
            echo "Kafka broker ${BROKER_ID} is not registered yet, retry ${i}..."
            sleep 5
          fi
        done
        echo "Failed to start Kafka broker ${BROKER_ID}." >&2
        exit 1
      }
      
      function install_and_configure_kafka_server() {
        # Find zookeeper list first, before attempting any installation.
        local zookeeper_client_port
        zookeeper_client_port=$(grep 'clientPort' /etc/zookeeper/conf/zoo.cfg |
          tail -n 1 |
          cut -d '=' -f 2)
      
        local zookeeper_list
        zookeeper_list=$(grep '^server\.' /etc/zookeeper/conf/zoo.cfg |
          cut -d '=' -f 2 |
          cut -d ':' -f 1 |
          sort |
          uniq |
          sed "s/$/:${zookeeper_client_port}/" |
          xargs echo |
          sed "s/ /,/g")
      
        if [[ -z "${zookeeper_list}" ]]; then
          # Didn't find zookeeper quorum in zoo.cfg, but possibly workers just didn't
          # bother to populate it. Check if YARN HA is configured.
          zookeeper_list=$(bdconfig get_property_value --configuration_file \
            /etc/hadoop/conf/yarn-site.xml \
            --name yarn.resourcemanager.zk-address 2>/dev/null)
        fi
      
        # If all attempts failed, error out.
        if [[ -z "${zookeeper_list}" ]]; then
          err 'Failed to find configured Zookeeper list; try "--num-masters=3" for HA'
        fi
      
        ZOOKEEPER_ADDRESS="${zookeeper_list%%,*}"
      
        # Install Kafka from Dataproc distro.
        install_apt_get kafka-server || dpkg -l kafka-server ||
          err 'Unable to install and find kafka-server.'
      
        mkdir -p /var/lib/kafka-logs
        chown kafka:kafka -R /var/lib/kafka-logs
      
        if [[ "${ROLE}" == "Master" ]]; then
          # For master nodes, broker ID starts from 10,000.
          if [[ "$(hostname)" == *-m ]]; then
            # non-HA
            BROKER_ID=10000
          else
            # HA
            BROKER_ID=$((10000 + $(hostname | sed 's/.*-m-\([0-9]*\)$/\1/g')))
          fi
        else
          # For worker nodes, broker ID is a random number generated less than 10000.
          # 10000 is choosen since the max broker ID allowed being set is 10000.
          BROKER_ID=$((RANDOM % 10000))
        fi
        sed -i 's|log.dirs=/tmp/kafka-logs|log.dirs=/var/lib/kafka-logs|' \
          "${KAFKA_PROP_FILE}"
        sed -i 's|^\(zookeeper\.connect=\).*|\1'${zookeeper_list}'|' \
          "${KAFKA_PROP_FILE}"
        sed -i 's,^\(broker\.id=\).*,\1'${BROKER_ID}',' \
          "${KAFKA_PROP_FILE}"
        echo -e '\nreserved.broker.max.id=100000' >>"${KAFKA_PROP_FILE}"
        echo -e '\ndelete.topic.enable=true' >>"${KAFKA_PROP_FILE}"
      
        if [[ "${KAFKA_ENABLE_JMX}" == "true" ]]; then
          sed -i '/kafka-run-class.sh/i export KAFKA_JMX_OPTS="-Dcom.sun.management.jmxremote=true -Dcom.sun.management.jmxremote.authenticate=false -Dcom.sun.management.jmxremote.ssl=false -Djava.rmi.server.hostname=localhost -Djava.net.preferIPv4Stack=true"' /usr/lib/kafka/bin/kafka-server-start.sh
          sed -i "/kafka-run-class.sh/i export JMX_PORT=${KAFKA_JMX_PORT}" /usr/lib/kafka/bin/kafka-server-start.sh
        fi
      
        wait_for_zookeeper
      
        # Start Kafka.
        service kafka-server restart
      
        wait_for_kafka
      }
      
      function install_kafka_python_package() {
        KAFKA_PYTHON_PACKAGE="kafka-python==2.0.2"
        if [[ "${INSTALL_KAFKA_PYTHON}" != "true" ]]; then
          return
        fi
      
        if [[ "$(echo "${DATAPROC_IMAGE_VERSION} > 2.0" | bc)" -eq 1 ]]; then
          /opt/conda/default/bin/pip install "${KAFKA_PYTHON_PACKAGE}" || { sleep 10; /opt/conda/default/bin/pip install "${KAFKA_PYTHON_PACKAGE}"; }
        else
          OS=$(. /etc/os-release && echo "${ID}")
          if [[ "${OS}" == "rocky" ]]; then
            yum install -y python2-pip
          else
            apt-get install -y python-pip
          fi
          pip2 install "${KAFKA_PYTHON_PACKAGE}" || { sleep 10; pip2 install "${KAFKA_PYTHON_PACKAGE}"; } || { sleep 10; pip install "${KAFKA_PYTHON_PACKAGE}"; }
        fi
      }
      
      function remove_old_backports {
        # This script uses 'apt-get update' and is therefore potentially dependent on
        # backports repositories which have been archived.  In order to mitigate this
        # problem, we will remove any reference to backports repos older than oldstable
      
        # https://github.com/GoogleCloudDataproc/initialization-actions/issues/1157
        oldstable=$(curl -s https://deb.debian.org/debian/dists/oldstable/Release | awk '/^Codename/ {print $2}');
        stable=$(curl -s https://deb.debian.org/debian/dists/stable/Release | awk '/^Codename/ {print $2}');
      
        matched_files="$(grep -rsil '\-backports' /etc/apt/sources.list*)"
        if [[ -n "$matched_files" ]]; then
          for filename in "$matched_files"; do
            grep -e "$oldstable-backports" -e "$stable-backports" "$filename" || \
              sed -i -e 's/^.*-backports.*$//' "$filename"
          done
        fi
      }
      
      function main() {
        OS=$(. /etc/os-release && echo "${ID}")
        if [[ ${OS} == debian ]] && [[ $(echo "${DATAPROC_IMAGE_VERSION} <= 2.1" | bc -l) == 1 ]]; then
          remove_old_backports
        fi
        recv_keys || err 'Unable to receive keys.'
        update_apt_get || err 'Unable to update packages lists.'
        install_kafka_python_package
      
        # Only run the installation on workers; verify zookeeper on master(s).
        if [[ "${ROLE}" == 'Master' ]]; then
          service zookeeper-server status ||
            err 'Required zookeeper-server not running on master!'
          if [[ "${RUN_ON_MASTER}" == "true" ]]; then
            # Run installation on masters.
            install_and_configure_kafka_server
          else
            # On master nodes, just install kafka command-line tools and libs but not
            # kafka-server.
            install_apt_get kafka ||
              err 'Unable to install kafka libraries on master!'
          fi
        else
          # Run installation on workers.
          install_and_configure_kafka_server
        fi
      }
      
      main
      

    2. kafka.sh 初始化操作脚本复制到 Cloud Storage 存储桶中。此脚本会在 Dataproc 集群上安装 Kafka。

      1. 打开 Cloud Shell,然后运行以下命令:

        gcloud storage cp gs://goog-dataproc-initialization-actions-REGION/kafka/kafka.sh gs://BUCKET_NAME/scripts/
        

        进行以下替换:

        • REGIONkafka.sh 存储在 Cloud Storage 中带有区域标记的公共存储桶。指定地理位置相近的 Compute Engine 区域(例如 us-central1)。
        • BUCKET_NAME - Cloud Storage 存储桶的名称。

    创建 Dataproc Kafka 集群

    1. 打开 Cloud Shell,然后运行以下 gcloud dataproc clusters create 命令,以创建用于安装 Kafka 和 ZooKeeper 组件的 Dataproc 高可用性集群

      gcloud dataproc clusters create KAFKA_CLUSTER \
          --project=PROJECT_ID \
          --region=REGION \
          --image-version=2.1-debian11 \
          --num-masters=3 \
          --enable-component-gateway \
          --initialization-actions=gs://BUCKET_NAME/scripts/kafka.sh
      

      注意:

      • KAFKA_CLUSTER:集群名称(在项目中必须是唯一的)。该名称必须以小写字母开头,最多可包含 51 个小写字母、数字和连字符,不能以连字符结尾。已删除集群的名称可以再次使用。
      • PROJECT_ID:要与此集群关联的项目。
      • REGION:集群所在的 Compute Engine 区域,例如 us-central1
        • 您可以添加可选的 --zone=ZONE 标志,以在指定区域内指定可用区,例如 us-central1-a。如果您未指定可用区,Dataproc 自动选择可用区功能会在指定区域内选择可用区。
      • --image-version:本教程建议使用 Dataproc 映像版本 2.1-debian11。注意:每个映像版本都包含一组预安装的组件,包括本教程中使用的 Hive 组件(请参阅支持的 Dataproc 映像版本)。
      • --num-master3 主节点用于创建高可用性集群。Kafka 所需的 Zookeeper 组件已预安装在高可用性集群上。
      • --enable-component-gateway:启用 Dataproc 组件网关
      • BUCKET_NAME:包含 /scripts/kafka.sh 初始化脚本的 Cloud Storage 存储桶的名称(请参阅将 Kafka 安装脚本复制到 Cloud Storage)。

    创建 Kafka custdata 主题

    如需在 Dataproc Kafka 集群上创建 Kafka 主题,请执行以下操作:

    1. 使用 SSH 实用程序在集群主虚拟机上打开终端窗口。

    2. 创建 Kafka custdata 主题。

      /usr/lib/kafka/bin/kafka-topics.sh \
          --bootstrap-server KAFKA_CLUSTER-w-0:9092 \
          --create --topic custdata
      

      注意:

      • KAFKA_CLUSTER:插入 Kafka 集群的名称。 -w-0:9092 表示在 worker-0 节点的端口 9092 上运行的 Kafka 代理。

      • 创建 custdata 主题后,您可以运行以下命令:

        # List all topics.
        /usr/lib/kafka/bin/kafka-topics.sh \
            --bootstrap-server KAFKA_CLUSTER-w-0:9092 \
            --list
        
        # Consume then display topic data. /usr/lib/kafka/bin/kafka-console-consumer.sh \     --bootstrap-server KAFKA_CLUSTER-w-0:9092 \     --topic custdata
        # Count the number of messages in the topic. /usr/lib/kafka/bin/kafka-run-class.sh kafka.tools.GetOffsetShell \     --broker-list KAFKA_CLUSTER-w-0:9092 \     --topic custdata
        # Delete topic. /usr/lib/kafka/bin/kafka-topics.sh \     --bootstrap-server KAFKA_CLUSTER-w-0:9092 \     --delete --topic custdata

    将内容发布到 Kafka custdata 主题

    以下脚本使用 kafka-console-producer.sh Kafka 工具生成 CSV 格式的虚构客户数据。

    1. 将该脚本复制并其粘贴到 Kafka 集群主节点上的 SSH 终端。按 <return> 运行脚本。

      for i in {1..10000}; do \
      custname="cust name${i}"
      uuid=$(dbus-uuidgen)
      age=$((45 + $RANDOM % 45))
      amount=$(echo "$(( $RANDOM % 99999 )).$(( $RANDOM % 99 ))")
      message="${uuid}:${custname},${age},${amount}"
      echo ${message}
      done | /usr/lib/kafka/bin/kafka-console-producer.sh \
      --broker-list KAFKA_CLUSTER-w-0:9092 \
      --topic custdata \
      --property "parse.key=true" \
      --property "key.separator=:"
      

      注意:

      • KAFKA_CLUSTER:Kafka 集群的名称。
    2. 运行以下 Kafka 命令,以确认 custdata 主题包含 10,000 条消息。

      /usr/lib/kafka/bin/kafka-run-class.sh kafka.tools.GetOffsetShell \
      --broker-list KAFKA_CLUSTER-w-0:9092 \
      --topic custdata
      

      注意:

      • KAFKA_CLUSTER:Kafka 集群的名称。

      预期输出:

      custdata:0:10000
      

    在 Cloud Storage 中创建 Hive 表

    创建 Hive 表以接收流式传输的 Kafka 主题数据。执行以下步骤,以在 Cloud Storage 存储桶中创建 cust_parquet (Parquet) 和 cust_orc (ORC) Hive 表。

    1. 在以下脚本中插入 BUCKET_NAME,将该脚本复制并粘贴到 Kafka 集群主节点上的 SSH 终端,然后按 <return> 以创建 ~/hivetables.hql(Hive 查询语言)脚本。

      您将在下一步中运行 ~/hivetables.hql 脚本,以在 Cloud Storage 存储桶中创建 Parquet 和 ORC Hive 表。

      cat > ~/hivetables.hql <<EOF
      drop table if exists cust_parquet;
      create external table if not exists cust_parquet
      (uuid string, custname string, age string, amount string)
      row format delimited fields terminated by ','
      stored as parquet
      location "gs://BUCKET_NAME/tables/cust_parquet";
      

      drop table if exists cust_orc; create external table if not exists cust_orc (uuid string, custname string, age string, amount string) row format delimited fields terminated by ',' stored as orc location "gs://BUCKET_NAME/tables/cust_orc"; EOF
    2. 在 Kafka 集群主节点上的 SSH 终端中,提交 ~/hivetables.hql Hive 作业,以在 Cloud Storage 存储桶中创建 cust_parquet (Parquet) 和 cust_orc (ORC) Hive 表。

      gcloud dataproc jobs submit hive \
          --cluster=KAFKA_CLUSTER \
          --region=REGION \
          -f ~/hivetables.hql
      

      注意:

      • Hive 组件已预安装在 Dataproc Kafka 集群上。如需查看最近发布的 2.1 映像中包含的 Hive 组件版本列表,请参阅 2.1.x 发布版本
      • KAFKA_CLUSTER:Kafka 集群的名称。
      • REGION:Kafka 集群所在的区域。

    将 Kafka custdata 流式传输到 Hive 表

    1. 在 Kafka 集群的主节点上的 SSH 终端中运行以下命令,以安装 kafka-python 库。需要使用 Kafka 客户端才能将 Kafka 主题数据流式传输到 Cloud Storage。
      pip install kafka-python
      
    2. 插入 BUCKET_NAME,将以下 PySpark 代码复制并粘贴到 Kafka 集群主节点上的 SSH 终端,然后按 <return> 以创建 streamdata.py 文件。

      该脚本会订阅 Kafka custdata 主题,然后将数据流式传输到 Cloud Storage 中的 Hive 表。输出格式(可以是 Parquet 或 ORC)作为参数传递到脚本中。

      cat > streamdata.py <<EOF
      #!/bin/python
      
      import sys
      from pyspark.sql.functions import *
      from pyspark.sql.types import *
      from pyspark.sql import SparkSession
      from kafka import KafkaConsumer
      
      def getNameFn (data): return data.split(",")[0]
      def getAgeFn  (data): return data.split(",")[1]
      def getAmtFn  (data): return data.split(",")[2]
      
      def main(cluster, outputfmt):
          spark = SparkSession.builder.appName("APP").getOrCreate()
          spark.sparkContext.setLogLevel("WARN")
          Logger = spark._jvm.org.apache.log4j.Logger
          logger = Logger.getLogger(__name__)
      
          rows = spark.readStream.format("kafka") \
          .option("kafka.bootstrap.servers", cluster+"-w-0:9092").option("subscribe", "custdata") \
          .option("startingOffsets", "earliest")\
          .load()
      
          getNameUDF = udf(getNameFn, StringType())
          getAgeUDF  = udf(getAgeFn,  StringType())
          getAmtUDF  = udf(getAmtFn,  StringType())
      
          logger.warn("Params passed in are cluster name: " + cluster + "  output format(sink): " + outputfmt)
      
          query = rows.select (col("key").cast("string").alias("uuid"),\
              getNameUDF      (col("value").cast("string")).alias("custname"),\
              getAgeUDF       (col("value").cast("string")).alias("age"),\
              getAmtUDF       (col("value").cast("string")).alias("amount"))
      
          writer = query.writeStream.format(outputfmt)\
                  .option("path","gs://BUCKET_NAME/tables/cust_"+outputfmt)\
                  .option("checkpointLocation", "gs://BUCKET_NAME/chkpt/"+outputfmt+"wr") \
              .outputMode("append")\
              .start()
      
          writer.awaitTermination()
      
      if __name__=="__main__":
          if len(sys.argv) < 2:
              print ("Invalid number of arguments passed ", len(sys.argv))
              print ("Usage: ", sys.argv[0], " cluster  format")
              print ("e.g.:  ", sys.argv[0], " <cluster_name>  orc")
              print ("e.g.:  ", sys.argv[0], " <cluster_name>  parquet")
          main(sys.argv[1], sys.argv[2])
      
      EOF
      
    3. 在 Kafka 集群主节点上的 SSH 终端中,运行 spark-submit 以将数据流式传输到 Cloud Storage 中的 Hive 表。

      1. 插入 KAFKA_CLUSTER 的名称和输出 FORMAT,将以下代码复制并粘贴到 Kafka 集群主节点上的 SSH 终端,然后按 <return> 运行代码,并将 Kafka custdata 数据以 Parquet 格式流式传输到 Cloud Storage 中的 Hive 表。

        spark-submit --packages \
        org.apache.spark:spark-streaming-kafka-0-10_2.12:3.1.3,org.apache.spark:spark-sql-kafka-0-10_2.12:3.1.3 \
            --conf spark.history.fs.gs.outputstream.type=FLUSHABLE_COMPOSITE \
            --conf spark.driver.memory=4096m \
            --conf spark.executor.cores=2 \
            --conf spark.executor.instances=2 \
            --conf spark.executor.memory=6144m \
            streamdata.py KAFKA_CLUSTER FORMAT
            

        注意:

        • KAFKA_CLUSTER:插入 Kafka 集群的名称。
        • FORMAT:指定 parquetorc 作为输出格式。您可以连续运行该命令,以将这两种格式的数据流式传输到 Hive 表中:例如,在第一次调用中,指定 parquet 以将 Kafka custdata 主题流式传输到 Hive Parquet 表;然后在第二次调用中,指定 orc 格式以将 custdata 流式传输到 Hive ORC 表。
    4. 标准输出在 SSH 终端中停止后(这表示所有 custdata 都已流式传输),请在 SSH 终端中按 <control-c> 以停止流程。

    5. 列出 Cloud Storage 中的 Hive 表。

      gcloud storage ls gs://BUCKET_NAME/tables/* --recursive
      

      注意:

      • BUCKET_NAME:插入包含 Hive 表的 Cloud Storage 存储桶的名称(请参阅创建 Hive 表)。

    查询流式传输的数据

    1. 在 Kafka 集群主节点上的 SSH 终端中,运行以下 hive 命令,以统计 Cloud Storage 的 Hive 表中流式传输的 Kafka custdata 消息的数量。

      hive -e "select count(1) from TABLE_NAME"
      

      注意:

      • TABLE_NAME:将 cust_parquetcust_orc 指定为 Hive 表名称。

      预期输出代码段:

    ...
    Status: Running (Executing on YARN cluster with App id application_....)
    
    ----------------------------------------------------------------------------------------------
            VERTICES      MODE        STATUS  TOTAL  COMPLETED  RUNNING  PENDING  FAILED  KILLED
    ----------------------------------------------------------------------------------------------
    Map 1 .......... container     SUCCEEDED      1          1        0        0       0       0
    Reducer 2 ...... container     SUCCEEDED      1          1        0        0       0       0
    ----------------------------------------------------------------------------------------------
    VERTICES: 02/02  [==========================>>] 100%  ELAPSED TIME: 9.89 s
    ----------------------------------------------------------------------------------------------
    OK
    10000
    Time taken: 21.394 seconds, Fetched: 1 row(s)
    

    清理

    删除项目

    1. In the Google Cloud console, go to the Manage resources page.

      Go to Manage resources

    2. In the project list, select the project that you want to delete, and then click Delete.
    3. In the dialog, type the project ID, and then click Shut down to delete the project.

    删除资源

    • In the Google Cloud console, go to the Cloud Storage Buckets page.

      Go to Buckets

    • Click the checkbox for the bucket that you want to delete.
    • To delete the bucket, click Delete, and then follow the instructions.
    • 删除 Kafka 集群:
      gcloud dataproc clusters delete KAFKA_CLUSTER \
          --region=${REGION}