CentOS+QT+KAFKA开发环境部署及测试

CentOS+QT+KAFKA开发环境部署及测试

本文档记录了在CentOS环境下通过QT开发KAFKA程序的步骤，关于CentOS中安装QT集成开发环境，不再赘述。此处默认是在QT编译环境已经完备的情况下，如何配置KAFKA的编译环境及测试实例演示。

• 安装librdkafka

librdkafa是一个开源的KAFKA客户端，由C/C++实现，提供了生产者、消费者、管理者的客户端，是一款稳定的、高性能的消息中间件。对于生产者每秒可发送百万量级的消息数，对于消费者每秒可以消费掉3百万量级的消息数。

1. 下载librdkafka源码库，地址：GitHub - edenhill/librdkafka: The Apache Kafka C/C++ library；
2. 解压到指定目录，cd ./librdkafka-master/

1. 执行命令：chmod 777 configure lds-gen.py
2. 执行命令：./configure

1. 执行指令：make

1. 执行指令：make install

（以管理员用户权限）

至此librdkafka安装完毕

• 安装 KAFKA服务器

1. 下载二进制安装包：Apache Download Mirrors，解压即可（免安装）
2. 下载JDK安装包：https://www.oracle.com/java/technologies/javase/javase-jdk8-downloads.html
3. 先安装JDK并配置系统环境变量，这是KAFKA服务运行的前提；
4. 启动zookeeper服务指令：.\windows\zookeeper-server-start.bat ..\config\zookeeper.properties

1. 启动kafka服务指令：.\windows\kafka-server-start.bat ..\config\server.properties

上述log中显示了kafka的信息，如密码、端口号等，默认端口号：

port = 9092

1. 创建主题指令：.\windows\kafka-topics.bat --zookeeper 127.0.0.1:2181 --create --topic demo-topic --partitions 2  --replication-factor 1

7、启动生产者指令：.\kafka-console-producer.bat --broker-list 127.0.0.1:9092 --topic demo

生产者发布数据

1. 启动消费者指令：.\kafka-console-consumer.bat --bootstrap-server localhost:9092 --topic demo --from-beginning

参数：--from-beginning表示游标从头开始，不加此参数则从当前位置开始。

.\windows\kafka-console-consumer.bat --bootstrap-server localhost:9092 --topic demo

消费者接收数据

• QT开发生产者和消费者程序

1. 生产者

创建基于控制台的Qt应用程序，键入源代码：

#include <QCoreApplication>

#include <iostream>

#include <string>

#include <cstdlib>

#include <cstdio>

#include <csignal>

#include <cstring>

#include <getopt.h>

#include "rdkafkacpp.h"

static bool run = true;

static void sigterm (int sig) {

    run = false;

}

class ExampleDeliveryReportCb : public RdKafka::DeliveryReportCb {

public:

    void dr_cb (RdKafka::Message &message) {

        std::cout << "Message delivery for (" << message.len() << " bytes): " <<message.errstr() << std::endl;

        if (message.key())

            std::cout << "Key: " << *(message.key()) << ";" << std::endl;

    }

};

class ExampleEventCb : public RdKafka::EventCb {

public:

    void event_cb (RdKafka::Event &event) {

        switch (event.type())

        {

        case RdKafka::Event::EVENT_ERROR:

            std::cerr << "ERROR (" << RdKafka::err2str(event.err()) << "): " <<event.str() << std::endl;

            if (event.err() == RdKafka::ERR__ALL_BROKERS_DOWN)

                run = false;

            break;

        case RdKafka::Event::EVENT_STATS:

            std::cerr << "\"STATS\": " << event.str() << std::endl;

            break;

        case RdKafka::Event::EVENT_LOG:

            fprintf(stderr, "LOG-%i-%s: %s\n",

                    event.severity(), event.fac().c_str(), event.str().c_str());

            break;

        default:

            std::cerr << "EVENT " << event.type() <<" (" << RdKafka::err2str(event.err()) << "): " <<event.str() << std::endl;

            break;

        }

    }

};

int main(int argc, char *argv[])

{

    QCoreApplication a(argc, argv);

//    std::string brokers = "localhost";

    std::string brokers = "192.168.3.103";

    std::string errstr;

//    std::string topic_str="test";

    std::string topic_str="demo";

    int32_t partition = RdKafka::Topic::PARTITION_UA;

    RdKafka::Conf *conf = RdKafka::Conf::create(RdKafka::Conf::CONF_GLOBAL);

    RdKafka::Conf *tconf = RdKafka::Conf::create(RdKafka::Conf::CONF_TOPIC);

    conf->set("bootstrap.servers", brokers, errstr);

    ExampleEventCb ex_event_cb;

    conf->set("event_cb", &ex_event_cb, errstr);

    signal(SIGINT, sigterm);

    signal(SIGTERM, sigterm);

    ExampleDeliveryReportCb ex_dr_cb;

    conf->set("dr_cb", &ex_dr_cb, errstr);

    RdKafka::Producer *producer = RdKafka::Producer::create(conf, errstr);

    if (!producer) {

        std::cerr << "Failed to create producer: " << errstr << std::endl;

        exit(1);

    }

    std::cout << "% Created producer " << producer->name() << std::endl;

    RdKafka::Topic *topic = RdKafka::Topic::create(producer, topic_str, tconf, errstr);

    if (!topic) {

        std::cerr << "Failed to create topic: " << errstr << std::endl;

        exit(1);

    }

    for (std::string line; run && std::getline(std::cin, line);) {

        if (line.empty()) {

            producer->poll(0);

            continue;

        }

        RdKafka::ErrorCode resp = producer->produce(topic, partition,

                                                    RdKafka::Producer::RK_MSG_COPY /* Copy payload */,

                                                    const_cast<char *>(line.c_str()), line.size(),

                                                    NULL, NULL);

        if (resp != RdKafka::ERR_NO_ERROR)

            std::cerr << "% Produce failed: " << RdKafka::err2str(resp) << std::endl;

        else

            std::cerr << "% Produced message (" << line.size() << " bytes)" << std::endl;

        producer->poll(0);

    }

    run = true;

    // 退出前处理完输出队列中的消息

    while (run && producer->outq_len() > 0) {

        std::cerr << "Waiting for " << producer->outq_len() << std::endl;

        producer->poll(1000);

    }

    delete conf;

    delete tconf;

    delete topic;

    delete producer;

    RdKafka::wait_destroyed(5000);

    return a.exec();

}

此时直接编译会报错，需做如下配置。

在.pro文件中添加引入librdkafka的库文件，

QMAKE_LFLAGS += -lrdkafka -lrdkafka++-lz -lpthread -lrt

#-lrdkafka等价于 LIBS += /usr/local/lib/librdkafka.so 

由于默认情况下librdkafka的头文件和库文件安装路径为：

/usr/local/include/librdkafka

/usr/local/lib  

此时需要在/etc/ld.so.conf文件中加入librdkafka.so所在的目录:/usr/local/lib/

方法一：

echo "/usr/local/lib" >> /etc/ld.so.conf
ldconfig

方法二：

vim /etc/ld.so.conf

ldconfig

然后在文件尾部添加librdkafka库的路径

编译出现如下错误

:-1: error: cannot find –lz

解决办法：

yum install zlib-devel

运行时错误日志：

LOG-3-FAIL: [thrd:DESKTOP-VC32RT7:9092/0]: DESKTOP-VC32RT7:9092/0: Failed to resolve 'DESKTOP-VC32RT7:9092': 未知的名称或服务 (after 10013ms in state CONNECT, 3 identical error(s) suppressed)

原因是默认情况下，kafka服务器不允许外部访问，打开./config/server.properties配置文件，更改如下：

1. 把原文件中31行的注释去掉

listeners=PLAINTEXT://:9092

1. 把36行的注释去掉，把advertised.listeners改为服务器的IP地址

修改前：advertised.listeners=PLAINTEXT://your.host.name:9092

修改后：advertised.listeners=PLAINTEXT://192.168.3.103:9092

然后重新启动zookeeper和kafka服务和消费者。在虚拟机上通过Qt编译运行生产者程序，可以在终端中输入消息内容，并回车，就可以在宿主即的消费者窗口打印输出消息了。

生产者发布消息

消费者订阅消息

1. 消费者

上例中使用的消费者为librdkafka中自带的消费者客户端，本例将自己创建消费者程序，订阅来自客户端的消息。

创建基于控制台的Qt应用程序，键入源代码：

#include <QCoreApplication>

#include <iostream>

#include <string>

#include <cstdlib>

#include <cstdio>

#include <csignal>

#include <cstring>

#include <sys/time.h>

#include <getopt.h>

#include <unistd.h>

#include "rdkafkacpp.h"

static bool run = true;

static bool exit_eof = true;

static int eof_cnt = 0;

static int partition_cnt = 0;

static int verbosity = 1;

static long msg_cnt = 0;

static int64_t msg_bytes = 0;

static void sigterm (int sig) {

    run = false;

}

class ExampleEventCb : public RdKafka::EventCb {

public:

    void event_cb (RdKafka::Event &event) {

        switch (event.type())

        {

        case RdKafka::Event::EVENT_ERROR:

            std::cerr << "ERROR (" << RdKafka::err2str(event.err()) << "): " <<event.str() << std::endl;

            if (event.err() == RdKafka::ERR__ALL_BROKERS_DOWN)

                run = false;

            break;

        case RdKafka::Event::EVENT_STATS:

            std::cerr << "\"STATS\": " << event.str() << std::endl;

            break;

        case RdKafka::Event::EVENT_LOG:

            fprintf(stderr, "LOG-%i-%s: %s\n",event.severity(), event.fac().c_str(), event.str().c_str());

            break;

        case RdKafka::Event::EVENT_THROTTLE:

            std::cerr << "THROTTLED: " << event.throttle_time() << "ms by "<<event.broker_name() << " id " << (int)event.broker_id() << std::endl;

            break;

        default:

            std::cerr << "EVENT " << event.type() <<" (" << RdKafka::err2str(event.err()) << "): " <<event.str() << std::endl;

            break;

        }

    }

};

void msg_consume(RdKafka::Message* message, void* opaque) {

    switch (message->err()) {

    case RdKafka::ERR__TIMED_OUT:

        //std::cerr << "RdKafka::ERR__TIMED_OUT"<<std::endl;

        break;

    case RdKafka::ERR_NO_ERROR:

        /* Real message */

        msg_cnt++;

        msg_bytes += message->len();

        if (verbosity >= 3)

            std::cerr << "Read msg at offset " << message->offset() << std::endl;

        RdKafka::MessageTimestamp ts;

        ts = message->timestamp();

        if (verbosity >= 2 &&ts.type != RdKafka::MessageTimestamp::MSG_TIMESTAMP_NOT_AVAILABLE) {

            std::string tsname = "?";

            if (ts.type == RdKafka::MessageTimestamp::MSG_TIMESTAMP_CREATE_TIME)

                tsname = "create time";

            else if (ts.type == RdKafka::MessageTimestamp::MSG_TIMESTAMP_LOG_APPEND_TIME)

                tsname = "log append time";

            std::cout << "Timestamp: " << tsname << " " << ts.timestamp << std::endl;

        }

        if (verbosity >= 2 && message->key()) {

            std::cout << "Key: " << *message->key() << std::endl;

        }

        if (verbosity >= 1) {

            printf("%.*s\n",static_cast<int>(message->len()),static_cast<const char *>(message->payload()));

        }

        break;

    case RdKafka::ERR__PARTITION_EOF:

        /* Last message */

        if (exit_eof && ++eof_cnt == partition_cnt) {

            std::cerr << "%% EOF reached for all " << partition_cnt <<" partition(s)" << std::endl;

            run = false;

        }

        break;

    case RdKafka::ERR__UNKNOWN_TOPIC:

        break;

    case RdKafka::ERR__UNKNOWN_PARTITION:

        std::cerr << "Consume failed: " << message->errstr() << std::endl;

        run = false;

        break;

    default:

        /* Errors */

        std::cerr << "Consume failed: " << message->errstr() << std::endl;

        run = false;

        break;

    }

}

class ExampleConsumeCb : public RdKafka::ConsumeCb {

public:

    void consume_cb (RdKafka::Message &msg, void *opaque) {

        msg_consume(&msg, opaque);

    }

};

int main(int argc, char *argv[])

{

    QCoreApplication a(argc, argv);

//    std::string brokers = "localhost";

    std::string brokers = "192.168.3.103";

    std::string errstr;

//    std::string topic_str="test";

    std::string topic_str="demo";

    std::vector<std::string> topics;

    std::string group_id="101";

    RdKafka::Conf *conf = RdKafka::Conf::create(RdKafka::Conf::CONF_GLOBAL);

    RdKafka::Conf *tconf = RdKafka::Conf::create(RdKafka::Conf::CONF_TOPIC);

    //group.id必须设置

    if (conf->set("group.id", group_id, errstr) != RdKafka::Conf::CONF_OK) {

        std::cerr << errstr << std::endl;

        exit(1);

    }

    topics.push_back(topic_str);

    //bootstrap.servers可以替换为metadata.broker.list

    conf->set("bootstrap.servers", brokers, errstr);

    ExampleConsumeCb ex_consume_cb;

    conf->set("consume_cb", &ex_consume_cb, errstr);

    ExampleEventCb ex_event_cb;

    conf->set("event_cb", &ex_event_cb, errstr);

    conf->set("default_topic_conf", tconf, errstr);

    signal(SIGINT, sigterm);

    signal(SIGTERM, sigterm);

    RdKafka::KafkaConsumer *consumer = RdKafka::KafkaConsumer::create(conf, errstr);

    if (!consumer) {

        std::cerr << "Failed to create consumer: " << errstr << std::endl;

        exit(1);

    }

    std::cout << "% Created consumer " << consumer->name() << std::endl;

    RdKafka::ErrorCode err = consumer->subscribe(topics);

    if (err) {

        std::cerr << "Failed to subscribe to " << topics.size() << " topics: " << RdKafka::err2str(err) << std::endl;

        exit(1);

    }

    while (run) {

    //5000毫秒未订阅到消息，触发RdKafka::ERR__TIMED_OUT

    RdKafka::Message *msg = consumer->consume(5000);

    msg_consume(msg, NULL);

    delete msg;

    }

    consumer->close();

    delete conf;

    delete tconf;

    delete consumer;

    std::cerr << "% Consumed " << msg_cnt << " messages ("<< msg_bytes << " bytes)" << std::endl;

    //应用退出之前等待rdkafka清理资源

    RdKafka::wait_destroyed(5000);

    return a.exec();

}

在.pro文件中添加librdkafka库路径，此处演示了另一种更常规的添加库路径的方式：

QT -= gui

CONFIG += c++11 console

CONFIG -= app_bundle

DEFINES += QT_DEPRECATED_WARNINGS

SOURCES += \

        main.cpp

# Default rules for deployment.

qnx: target.path = /tmp/$${TARGET}/bin

else: unix:!android: target.path = /opt/$${TARGET}/bin

!isEmpty(target.path): INSTALLS += target

INCLUDEPATH += /usr/local/include/librdkafka

LIBS += -L/usr/local/lib -lrdkafka

LIBS += -L/usr/local/lib -lrdkafka++

编译并运行消费者程序，

生产者发布消息：

消费者订阅消息：

参考来源：

1. https://blog.csdn.net/caoshangpa/article/details/79786100?utm_medium=distribute.pc_relevant.none-task-blog-2~default~BlogCommendFromMachineLearnPai2~default-5.base&depth_1-utm_source=distribute.pc_relevant.none-task-blog-2~default~BlogCommendFromMachineLearnPai2~default-5.base
2. https://www.cnblogs.com/alvingofast/p/kafka_deployment_on_windows.html