c语言使用librdkafka库实现kafka的生产和消费实例

关于librdkafka库的介绍，可以参考kafka的c/c++高性能客户端librdkafka简介，本文使用librdkafka库来进行kafka的简单的生产、消费

一、producer

librdkafka进行kafka生产操作的大致步骤如下：

1、创建kafka配置

rd_kafka_conf_t *rd_kafka_conf_new (void)

2、配置kafka各项参数

rd_kafka_conf_res_t rd_kafka_conf_set (rd_kafka_conf_t *conf,
const char *name,
const char *value,
char *errstr, size_t errstr_size)

3、设置发送回调函数

void rd_kafka_conf_set_dr_msg_cb (rd_kafka_conf_t *conf,
void (*dr_msg_cb) (rd_kafka_t *rk,
const rd_kafka_message_t *
rkmessage,
void *opaque))

4、创建producer实例

rd_kafka_t *rd_kafka_new (rd_kafka_type_t type, rd_kafka_conf_t *conf,char *errstr, size_t errstr_size)

5、实例化topic

rd_kafka_topic_t *rd_kafka_topic_new (rd_kafka_t *rk, const char *topic, rd_kafka_topic_conf_t *conf)

6、异步调用将消息发送到指定的topic

int rd_kafka_produce (rd_kafka_topic_t *rkt, int32_t partition,
int msgflags,
void *payload, size_t len,
const void *key, size_t keylen,
void *msg_opaque)

7、阻塞等待消息发送完成

int rd_kafka_poll (rd_kafka_t *rk, int timeout_ms)

8、等待完成producer请求完成

rd_kafka_resp_err_t rd_kafka_flush (rd_kafka_t *rk, int timeout_ms)

9、销毁topic

void rd_kafka_topic_destroy (rd_kafka_topic_t *app_rkt)

10、销毁producer实例

void rd_kafka_destroy (rd_kafka_t *rk)

完整代码如下my_producer.c：

#include <stdio.h>
#include <signal.h>
#include <string.h>

#include "../src/rdkafka.h"

static int run = 1;

static void stop(int sig){
run = 0;
fclose(stdin);
}

/*
每条消息调用一次该回调函数，说明消息是传递成功(rkmessage->err == RD_KAFKA_RESP_ERR_NO_ERROR)
还是传递失败(rkmessage->err != RD_KAFKA_RESP_ERR_NO_ERROR)
该回调函数由rd_kafka_poll()触发，在应用程序的线程上执行
*/
static void dr_msg_cb(rd_kafka_t *rk,
const rd_kafka_message_t *rkmessage, void *opaque){
if(rkmessage->err)
fprintf(stderr, "%% Message delivery failed: %s\n",
rd_kafka_err2str(rkmessage->err));
else
fprintf(stderr,
"%% Message delivered (%zd bytes, "
"partition %"PRId32")\n",
rkmessage->len, rkmessage->partition);
/* rkmessage被librdkafka自动销毁*/
}

int main(int argc, char **argv){
rd_kafka_t *rk;            /*Producer instance handle*/
rd_kafka_topic_t *rkt;     /*topic对象*/
rd_kafka_conf_t *conf;     /*临时配置对象*/
char errstr[512];
char buf[512];
const char *brokers;
const char *topic;

if(argc != 3){
fprintf(stderr, "%% Usage: %s <broker> <topic>\n", argv[0]);
return 1;
}

brokers = argv[1];
topic = argv[2];

/* 创建一个kafka配置占位 */
conf = rd_kafka_conf_new();

/*创建broker集群*/
if (rd_kafka_conf_set(conf, "bootstrap.servers", brokers, errstr,
sizeof(errstr)) != RD_KAFKA_CONF_OK){
fprintf(stderr, "%s\n", errstr);
return 1;
}

/*设置发送报告回调函数，rd_kafka_produce()接收的每条消息都会调用一次该回调函数
*应用程序需要定期调用rd_kafka_poll()来服务排队的发送报告回调函数*/
rd_kafka_conf_set_dr_msg_cb(conf, dr_msg_cb);

/*创建producer实例
rd_kafka_new()获取conf对象的所有权,应用程序在此调用之后不得再次引用它*/
rk = rd_kafka_new(RD_KAFKA_PRODUCER, conf, errstr, sizeof(errstr));
if(!rk){
fprintf(stderr, "%% Failed to create new producer:%s\n", errstr);
return 1;
}

/*实例化一个或多个topics(`rd_kafka_topic_t`)来提供生产或消费，topic
对象保存topic特定的配置，并在内部填充所有可用分区和leader brokers，*/
rkt = rd_kafka_topic_new(rk, topic, NULL);
if (!rkt){
fprintf(stderr, "%% Failed to create topic object: %s\n",
rd_kafka_err2str(rd_kafka_last_error()));
rd_kafka_destroy(rk);
return 1;
}

/*用于中断的信号*/
signal(SIGINT, stop);

fprintf(stderr,
"%% Type some text and hit enter to produce message\n"
"%% Or just hit enter to only serve delivery reports\n"
"%% Press Ctrl-C or Ctrl-D to exit\n");

while(run && fgets(buf, sizeof(buf), stdin)){
size_t len = strlen(buf);

if(buf[len-1] == '\n')
buf[--len] = '\0';

if(len == 0){
/*轮询用于事件的kafka handle，
事件将导致应用程序提供的回调函数被调用
第二个参数是最大阻塞时间，如果设为0，将会是非阻塞的调用*/
rd_kafka_poll(rk, 0);
continue;
}

retry:
/*Send/Produce message.
这是一个异步调用，在成功的情况下，只会将消息排入内部producer队列，
对broker的实际传递尝试由后台线程处理，之前注册的传递回调函数(dr_msg_cb)
用于在消息传递成功或失败时向应用程序发回信号*/
if (rd_kafka_produce(
/* Topic object */
rkt,
/*使用内置的分区来选择分区*/
RD_KAFKA_PARTITION_UA,
/*生成payload的副本*/
RD_KAFKA_MSG_F_COPY,
/*消息体和长度*/
buf, len,
/*可选键及其长度*/
NULL, 0,
NULL) == -1){
fprintf(stderr,
"%% Failed to produce to topic %s: %s\n",
rd_kafka_topic_name(rkt),
rd_kafka_err2str(rd_kafka_last_error()));

if (rd_kafka_last_error() == RD_KAFKA_RESP_ERR__QUEUE_FULL){
/*如果内部队列满，等待消息传输完成并retry,
内部队列表示要发送的消息和已发送或失败的消息，
内部队列受限于queue.buffering.max.messages配置项*/
rd_kafka_poll(rk, 1000);
goto retry;
}
}else{
fprintf(stderr, "%% Enqueued message (%zd bytes) for topic %s\n",
len, rd_kafka_topic_name(rkt));
}

/*producer应用程序应不断地通过以频繁的间隔调用rd_kafka_poll()来为
传送报告队列提供服务。在没有生成消息以确定先前生成的消息已发送了其
发送报告回调函数(和其他注册过的回调函数)期间，要确保rd_kafka_poll()
仍然被调用*/
rd_kafka_poll(rk, 0);
}

fprintf(stderr, "%% Flushing final message.. \n");
/*rd_kafka_flush是rd_kafka_poll()的抽象化，
等待所有未完成的produce请求完成，通常在销毁producer实例前完成
以确保所有排列中和正在传输的produce请求在销毁前完成*/
rd_kafka_flush(rk, 10*1000);

/* Destroy topic object */
rd_kafka_topic_destroy(rkt);

/* Destroy the producer instance */
rd_kafka_destroy(rk);

return 0;
}

二、consumer

librdkafka进行kafka消费操作的大致步骤如下：

1、创建kafka配置

rd_kafka_conf_t *rd_kafka_conf_new (void)

2、创建kafka topic的配置

rd_kafka_topic_conf_t *rd_kafka_topic_conf_new (void)

3、配置kafka各项参数

rd_kafka_conf_res_t rd_kafka_conf_set (rd_kafka_conf_t *conf,
const char *name,
const char *value,
char *errstr, size_t errstr_size)

4、配置kafka topic各项参数

rd_kafka_conf_res_t rd_kafka_topic_conf_set (rd_kafka_topic_conf_t *conf,
const char *name,
const char *value,
char *errstr, size_t errstr_size)

5、创建consumer实例

rd_kafka_t *rd_kafka_new (rd_kafka_type_t type, rd_kafka_conf_t *conf, char *errstr, size_t errstr_size)

6、为consumer实例添加brokerlist

int rd_kafka_brokers_add (rd_kafka_t *rk, const char *brokerlist)

7、开启consumer订阅

rd_kafka_subscribe (rd_kafka_t *rk, const rd_kafka_topic_partition_list_t *topics)

8、轮询消息或事件，并调用回调函数

rd_kafka_message_t *rd_kafka_consumer_poll (rd_kafka_t *rk,int timeout_ms)

9、关闭consumer实例

rd_kafka_resp_err_t rd_kafka_consumer_close (rd_kafka_t *rk)

10、释放topic list资源

rd_kafka_topic_partition_list_destroy (rd_kafka_topic_partition_list_t *rktparlist)

11、销毁consumer实例

void rd_kafka_destroy (rd_kafka_t *rk)

12、等待consumer对象的销毁

int rd_kafka_wait_destroyed (int timeout_ms)

完整代码如下my_consumer.c

#include <string.h>
#include <stdlib.h>
#include <syslog.h>
#include <signal.h>
#include <error.h>
#include <getopt.h>

#include "../src/rdkafka.h"

static int run = 1;
//`rd_kafka_t`自带一个可选的配置API，如果没有调用API，Librdkafka将会使用CONFIGURATION.md中的默认配置。
static rd_kafka_t *rk;
static rd_kafka_topic_partition_list_t *topics;

static void stop (int sig) {
if (!run)
exit(1);
run = 0;
fclose(stdin); /* abort fgets() */
}

static void sig_usr1 (int sig) {
rd_kafka_dump(stdout, rk);
}

/**
* 处理并打印已消费的消息
*/
static void msg_consume (rd_kafka_message_t *rkmessage,
void *opaque) {
if (rkmessage->err) {
if (rkmessage->err == RD_KAFKA_RESP_ERR__PARTITION_EOF) {
fprintf(stderr,
"%% Consumer reached end of %s [%"PRId32"] "
"message queue at offset %"PRId64"\n",
rd_kafka_topic_name(rkmessage->rkt),
rkmessage->partition, rkmessage->offset);

return;
}

if (rkmessage->rkt)
fprintf(stderr, "%% Consume error for "
"topic \"%s\" [%"PRId32"] "
"offset %"PRId64": %s\n",
rd_kafka_topic_name(rkmessage->rkt),
rkmessage->partition,
rkmessage->offset,
rd_kafka_message_errstr(rkmessage));
else
fprintf(stderr, "%% Consumer error: %s: %s\n",
rd_kafka_err2str(rkmessage->err),
rd_kafka_message_errstr(rkmessage));

if (rkmessage->err == RD_KAFKA_RESP_ERR__UNKNOWN_PARTITION ||
rkmessage->err == RD_KAFKA_RESP_ERR__UNKNOWN_TOPIC)
run = 0;
return;
}

fprintf(stdout, "%% Message (topic %s [%"PRId32"], "
"offset %"PRId64", %zd bytes):\n",
rd_kafka_topic_name(rkmessage->rkt),
rkmessage->partition,
rkmessage->offset, rkmessage->len);

if (rkmessage->key_len) {
printf("Key: %.*s\n",
(int)rkmessage->key_len, (char *)rkmessage->key);
}

printf("%.*s\n",
(int)rkmessage->len, (char *)rkmessage->payload);

}

/*
init all configuration of kafka
*/
int initKafka(char *brokers, char *group,char *topic){
rd_kafka_conf_t *conf;
rd_kafka_topic_conf_t *topic_conf;
rd_kafka_resp_err_t err;
char tmp[16];
char errstr[512];

/* Kafka configuration */
conf = rd_kafka_conf_new();

//quick termination
snprintf(tmp, sizeof(tmp), "%i", SIGIO);
rd_kafka_conf_set(conf, "internal.termination.signal", tmp, NULL, 0);

//topic configuration
topic_conf = rd_kafka_topic_conf_new();

/* Consumer groups require a group id */
if (!group)
group = "rdkafka_consumer_example";
if (rd_kafka_conf_set(conf, "group.id", group,
errstr, sizeof(errstr)) !=
RD_KAFKA_CONF_OK) {
fprintf(stderr, "%% %s\n", errstr);
return -1;
}

/* Consumer groups always use broker based offset storage */
if (rd_kafka_topic_conf_set(topic_conf, "offset.store.method",
"broker",
errstr, sizeof(errstr)) !=
RD_KAFKA_CONF_OK) {
fprintf(stderr, "%% %s\n", errstr);
return -1;
}

/* Set default topic config for pattern-matched topics. */
rd_kafka_conf_set_default_topic_conf(conf, topic_conf);

//实例化一个顶级对象rd_kafka_t作为基础容器，提供全局配置和共享状态
rk = rd_kafka_new(RD_KAFKA_CONSUMER, conf, errstr, sizeof(errstr));
if(!rk){
fprintf(stderr, "%% Failed to create new consumer:%s\n", errstr);
return -1;
}

//Librdkafka需要至少一个brokers的初始化list
if (rd_kafka_brokers_add(rk, brokers) == 0){
fprintf(stderr, "%% No valid brokers specified\n");
return -1;
}

//重定向 rd_kafka_poll()队列到consumer_poll()队列
rd_kafka_poll_set_consumer(rk);

//创建一个Topic+Partition的存储空间(list/vector)
topics = rd_kafka_topic_partition_list_new(1);
//把Topic+Partition加入list
rd_kafka_topic_partition_list_add(topics, topic, -1);
//开启consumer订阅，匹配的topic将被添加到订阅列表中
if((err = rd_kafka_subscribe(rk, topics))){
fprintf(stderr, "%% Failed to start consuming topics: %s\n", rd_kafka_err2str(err));
return -1;
}

return 1;
}

int main(int argc, char **argv){
char *brokers = "localhost:9092";
char *group = NULL;
char *topic = NULL;

int opt;
rd_kafka_resp_err_t err;

while ((opt = getopt(argc, argv, "g:b:t:qd:eX:As:DO")) != -1){
switch (opt) {
case 'b':
brokers = optarg;
break;
case 'g':
group = optarg;
break;
case 't':
topic = optarg;
break;
default:
break;
}
}

signal(SIGINT, stop);
signal(SIGUSR1, sig_usr1);

if(!initKafka(brokers, group, topic)){
fprintf(stderr, "kafka server initialize error\n");
}else{
while(run){
rd_kafka_message_t *rkmessage;
/*-轮询消费者的消息或事件，最多阻塞timeout_ms
-应用程序应该定期调用consumer_poll()，即使没有预期的消息，以服务
所有排队等待的回调函数，当注册过rebalance_cb，该操作尤为重要，
因为它需要被正确地调用和处理以同步内部消费者状态 */
rkmessage = rd_kafka_consumer_poll(rk, 1000);
if(rkmessage){
msg_consume(rkmessage, NULL);
/*释放rkmessage的资源，并把所有权还给rdkafka*/
rd_kafka_message_destroy(rkmessage);
}
}
}

done:
/*此调用将会阻塞，直到consumer撤销其分配，调用rebalance_cb(如果已设置)，
commit offset到broker,并离开consumer group
最大阻塞时间被设置为session.timeout.ms
*/
err = rd_kafka_consumer_close(rk);
if(err){
fprintf(stderr, "%% Failed to close consumer: %s\n", rd_kafka_err2str(err));
}else{
fprintf(stderr, "%% Consumer closed\n");
}

//释放topics list使用的所有资源和它自己
rd_kafka_topic_partition_list_destroy(topics);

//destroy kafka handle
rd_kafka_destroy(rk);

run = 5;
//等待所有rd_kafka_t对象销毁，所有kafka对象被销毁，返回0，超时返回-1
while(run-- > 0 && rd_kafka_wait_destroyed(1000) == -1){
printf("Waiting for librdkafka to decommission\n");
}
if(run <= 0){
//dump rdkafka内部状态到stdout流
rd_kafka_dump(stdout, rk);
}

return 0;
}

在linux下编译producer和consumer的代码：

gcc my_producer.c -o my_producer  -lrdkafka -lz -lpthread -lrt
gcc my_consumer.c -o my_consumer  -lrdkafka -lz -lpthread -lrt

在运行my_producer或my_consumer时可能会报错"error while loading shared libraries
xxx.so", 此时需要在/etc/ld.so.conf中加入xxx.so所在的目录

在本地启动一个简单的kafka服务，设置broker集群为localhost：9092并创建一个叫“test_topic”的topic

启动方式可参考 kafka0.8.2集群的环境搭建并实现基本的生产消费

启动consumer：



启动producer，并发送一条数据“hello world”：



consumer处成功收到producer发送的“hello world”：