weixin_34332905
weixin_34332905

关于kafka中consumer subscribe与asssign的理解

结论先行

凡是 1、具有相同group.id的consumer；2、将offset提交到kafka保存（不分手动和自动提交） 的consumer共享/共同维护 kafka中该group.id对应的offset。

那么就会有下面的影响：

1、broker中保存的是该group.id下所有consumer中最新提交的offset。

2、当有新的consumer启动的时候，其具体的流程应该是：

2-1、去broker查看是否有该`group.id`的存储信息，有的话从该`offset`开始消费，没有的话根据`auto.offset.reset`项配置。

2-2、consumer开始从2-1中的offset开始消费，并不断向broker提交offset更新broker中保存的该`group.id`对应的offset值。但是consumer在正常消费过程中就不再从kafka拉取最新的offset，只是向kafka不断的更新offset。
复制代码

3、强烈建议不要混用assign和subscribe方法。

4、seek函数可以显示指定offset，为粒度，单纯的seek并不会提交到到broker，需要同consumer commit一次才能生效，在enable.auto.commit=true的情况下是在poll的时候进行commit。

所以：

强烈推荐只使用subscribeapi.

高端用户慎用assignapi.

测试

assign

两个consumer相同配置，均采用assign方法，查看seek是否会相互影响。

首先，两个consumer之间会有影响，共享一套offset。

public void bothAssign() {
        KafkaConsumer consumer_a = new KafkaConsumer(props);
        KafkaConsumer consumer_b = new KafkaConsumer(props);
        ConsumerRecords msgs;

        TopicPartition tp = new TopicPartition(topic, 0);
        // 正常消费 ok 互不影响
        consumer_a.assign(Arrays.asList(tp));
        consumer_b.assign(Arrays.asList(tp));
        int a=0, b=0;

        while (a++ < 10) {
            msgs = consumer_a.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" A : offset {}, value {}.", record.offset(), record.value());
            }
        }

        while (b++ < 10) {
            msgs = consumer_b.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" B : offset {}, value {}.", record.offset(), record.value());
            }
        }
    }

# output
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,306(290) --> [main] org.apache.kafka.common.utils.AppInfoParser$AppInfo.(AppInfoParser.java:109): Kafka version : 1.0.1  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,307(291) --> [main] org.apache.kafka.common.utils.AppInfoParser$AppInfo.(AppInfoParser.java:110): Kafka commitId : c0518aa65f25317e  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,622(606) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-1, groupId=seekTest24] Discovered group coordinator g1-bdp-cdhtest-01.dns.guazi.com:9092 (id: 2147483233 rack: null)  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,764(748) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 24, value 24.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,766(750) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 25, value 25.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,767(751) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 26, value 26.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,768(752) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 27, value 27.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,769(753) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 28, value 28.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,770(754) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 29, value 29.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,771(755) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 30, value 30.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,772(756) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 31, value 31.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,774(758) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:44):  A : offset 32, value 32.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,872(856) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest24] Discovered group coordinator g1-bdp-cdhtest-01.dns.guazi.com:9092 (id: 2147483233 rack: null)  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,914(898) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 32, value 32.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,914(898) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 33, value 33.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,915(899) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 34, value 34.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,915(899) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 35, value 35.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,915(899) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 36, value 36.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,915(899) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 37, value 37.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,916(900) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 38, value 38.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,917(901) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 39, value 39.  
2018-12-20 13:43:22 [INFO ] 2018-12-20 13:43:22,917(901) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:51):  B : offset 40, value 40.  
复制代码

其次，两者之间seek互不影响，会从最后一次seek的值处开始消费。

# code
public void bothAssign() {
        KafkaConsumer consumer_a = new KafkaConsumer(props);
        KafkaConsumer consumer_b = new KafkaConsumer(props);
        ConsumerRecords msgs;

        TopicPartition tp = new TopicPartition(topic, 0);
        consumer_a.assign(Arrays.asList(tp));
        consumer_b.assign(Arrays.asList(tp));
        int a=0, b=0;

        while (a++ < 10) {
            msgs = consumer_a.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" A : offset {}, value {}.", record.offset(), record.value());
            }
        }

        while (b++ < 10) {
            msgs = consumer_b.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" B : offset {}, value {}.", record.offset(), record.value());
            }
        }

        logger.info("===================================");

        consumer_a.seek(tp, 33);
        consumer_b.seek(tp, 66);
        while (a++ < 20) {
            msgs = consumer_a.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" A : offset {}, value {}.", record.offset(), record.value());
            }
        }

        while (b++ < 20) {
            msgs = consumer_b.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" B : offset {}, value {}.", record.offset(), record.value());
            }
        }
    }
# output 
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,016(243) --> [main] org.apache.kafka.common.utils.AppInfoParser$AppInfo.(AppInfoParser.java:109): Kafka version : 1.0.1  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,016(243) --> [main] org.apache.kafka.common.utils.AppInfoParser$AppInfo.(AppInfoParser.java:110): Kafka commitId : c0518aa65f25317e  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,300(527) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-1, groupId=seekTest24] Discovered group coordinator g1-bdp-cdhtest-01.dns.guazi.com:9092 (id: 2147483233 rack: null)  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,408(635) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 32, value 32.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,409(636) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 33, value 33.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,411(638) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 34, value 34.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,412(639) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 35, value 35.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,413(640) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 36, value 36.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,414(641) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 37, value 37.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,414(641) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 38, value 38.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,415(642) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 39, value 39.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,415(642) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:43):  A : offset 40, value 40.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,494(721) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest24] Discovered group coordinator g1-bdp-cdhtest-01.dns.guazi.com:9092 (id: 2147483233 rack: null)  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,573(800) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 39, value 39.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,574(801) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 40, value 40.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,575(802) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 41, value 41.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,575(802) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 42, value 42.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,576(803) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 43, value 43.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,576(803) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 44, value 44.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,578(805) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 45, value 45.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,579(806) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 46, value 46.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,580(807) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:50):  B : offset 47, value 47.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,580(807) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:54): ===================================  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,598(825) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 66, value 66.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,599(826) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 67, value 67.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,600(827) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 68, value 68.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,602(829) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 69, value 69.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,604(831) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 70, value 70.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,605(832) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 71, value 71.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,606(833) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 72, value 72.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,606(833) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 73, value 73.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,608(835) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:61):  A : offset 74, value 74.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,714(941) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:68):  B : offset 70, value 70.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,715(942) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:68):  B : offset 71, value 71.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,716(943) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:68):  B : offset 72, value 72.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,721(948) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:68):  B : offset 73, value 73.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,722(949) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:68):  B : offset 74, value 74.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,722(949) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:68):  B : offset 75, value 75.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,723(950) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:68):  B : offset 76, value 76.  
2018-12-20 13:45:24 [INFO ] 2018-12-20 13:45:24,723(950) --> [main] com.songys.kafka.kafkaSeekTest.bothAssign(kafkaSeekTest.java:68):  B : offset 77, value 77.  
复制代码

说明，两者具有相同的client.id的情况下会报错。

subscribe

两个consumer相同配置，均采用subscribe方法，。

首先，两个consumer之间共享一套offset，互相影响，不能同时对同一个topic-partition进行seek。

#subscribe单并未poll之前进行seek则会报错.
Exception in thread "main" java.lang.IllegalStateException: No current assignment for partition seek_test_sys-0
	at org.apache.kafka.clients.consumer.internals.SubscriptionState.assignedState(SubscriptionState.java:269)
	at org.apache.kafka.clients.consumer.internals.SubscriptionState.seek(SubscriptionState.java:294)
	at org.apache.kafka.clients.consumer.KafkaConsumer.seek(KafkaConsumer.java:1339)
	at com.songys.kafka.kafkaSeekTest.bothSubscribe(kafkaSeekTest.java:90)
	at com.songys.kafka.kafkaSeekTest.main(kafkaSeekTest.java:112)
复制代码

assign和subscribe

首先，assign会去影响subscribe，反过来也是，所以不要混用：

public void assignAndSubscribe() {
        KafkaConsumer consumer_e = new KafkaConsumer(props);
        KafkaConsumer consumer_f = new KafkaConsumer(props);
        ConsumerRecords msgs;
        TopicPartition tp = new TopicPartition(topic, 0);

        consumer_e.assign(Arrays.asList(tp));
        consumer_f.subscribe(Arrays.asList(topic));
        int e=0, f=0;

        while (e++ < 10) {
            msgs = consumer_e.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" E : offset {}, value {}.", record.offset(), record.value());
            }
        }

        while (f++ < 10) {
            msgs = consumer_f.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" F : offset {}, value {}.", record.offset(), record.value());
            }
        }

    }

# output
2018-12-20 12:10:21 [INFO ] 2018-12-20 12:10:21,717(358) --> [main] org.apache.kafka.common.utils.AppInfoParser$AppInfo.(AppInfoParser.java:109): Kafka version : 1.0.1  
2018-12-20 12:10:21 [INFO ] 2018-12-20 12:10:21,717(358) --> [main] org.apache.kafka.common.utils.AppInfoParser$AppInfo.(AppInfoParser.java:110): Kafka commitId : c0518aa65f25317e  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,244(885) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-1, groupId=seekTest19] Discovered group coordinator g1-bdp-cdhtest-01.dns.guazi.com:9092 (id: 2147483233 rack: null)  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,422(1063) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 14, value 14.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,423(1064) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 15, value 15.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,424(1065) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 16, value 16.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,426(1067) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 17, value 17.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,426(1067) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 18, value 18.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,427(1068) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 19, value 19.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,428(1069) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 20, value 20.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,429(1070) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 21, value 21.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,430(1071) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 22, value 22.  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,564(1205) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest19] Discovered group coordinator g1-bdp-cdhtest-01.dns.guazi.com:9092 (id: 2147483233 rack: null)  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,571(1212) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest19] Revoking previously assigned partitions []  
2018-12-20 12:10:22 [INFO ] 2018-12-20 12:10:22,572(1213) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:336): [Consumer clientId=consumer-2, groupId=seekTest19] (Re-)joining group  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,644(4285) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest19] Successfully joined group with generation 3  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,645(4286) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest19] Setting newly assigned partitions [seek_test_sys-0]  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,767(4408) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:150):  F : offset 22, value 22.  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,768(4409) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:150):  F : offset 23, value 23.  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,769(4410) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:150):  F : offset 24, value 24.  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,770(4411) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:150):  F : offset 25, value 25.  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,770(4411) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:150):  F : offset 26, value 26.  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,771(4412) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:150):  F : offset 27, value 27.  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,772(4413) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:150):  F : offset 28, value 28.  
2018-12-20 12:10:25 [INFO ] 2018-12-20 12:10:25,772(4413) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:150):  F : offset 29, value 29.  
复制代码

consumer-subscribe 正常消费，然后seek，从seek处开始消费(必须要消费一次，来更新offset)。consumer-assign会从最后的offset处继续消费。

但是，consumer-subscribe的seek之后，使用consumer-subscribe则从seek的位置开始消费，正常。

但是如果consumer-subscribe的seek之后直接使用consumer-assign进行消费不会从seek处开始消费，因为没有提交上去。

consumer-assign正常消费，然后seek，从seek处开始消费(必须要消费一次，来更新offset)，consumer-subscribe从最新的offset开始消费。

但是，consumer-assign消费一段，consumer-subscribe会从其后面继续消费，这时候使用consumer-assign进行seek，其本身会从seek处开始进行消费，但是会出现问题，可能是例子写的不好，造成commit失败，终归到底是维护一个consumer-group的offset，所以肯定是应该有影响的。

# code
public void assignAndSubscribe() {
        KafkaConsumer consumer_e = new KafkaConsumer(props);
        KafkaConsumer consumer_f = new KafkaConsumer(props);
        ConsumerRecords msgs;
        TopicPartition tp = new TopicPartition(topic, 0);

        consumer_e.assign(Arrays.asList(tp));
        consumer_f.subscribe(Arrays.asList(topic));
        int e=0, f=0;

        while (e++ < 10) {
            msgs = consumer_e.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" E : offset {}, value {}.", record.offset(), record.value());
            }
        }

        while (f++ < 10) {
            msgs = consumer_f.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" F : offset {}, value {}.", record.offset(), record.value());
            }
        }

        consumer_e.seek(tp, 66);

        logger.info("seek");

        while (e++ < 20) {
            msgs = consumer_e.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" E : offset {}, value {}.", record.offset(), record.value());
            }
        }

        while (f++ < 20) {
            msgs = consumer_f.poll(100L);
            for (ConsumerRecord record : msgs) {
                logger.info(" F : offset {}, value {}.", record.offset(), record.value());
            }
        }
    }
# output 
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,066(322) --> [main] org.apache.kafka.common.utils.AppInfoParser$AppInfo.(AppInfoParser.java:109): Kafka version : 1.0.1  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,066(322) --> [main] org.apache.kafka.common.utils.AppInfoParser$AppInfo.(AppInfoParser.java:110): Kafka commitId : c0518aa65f25317e  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,504(760) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-1, groupId=seekTest24] Discovered group coordinator g1-bdp-cdhtest-01.dns.guazi.com:9092 (id: 2147483233 rack: null)  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,625(881) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 0, value 0.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,630(886) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 1, value 1.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,644(900) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 2, value 2.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,662(918) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 3, value 3.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,667(923) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 4, value 4.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,672(928) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 5, value 5.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,674(930) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 6, value 6.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,679(935) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 7, value 7.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,681(937) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:142):  E : offset 8, value 8.  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,709(965) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest24] Discovered group coordinator g1-bdp-cdhtest-01.dns.guazi.com:9092 (id: 2147483233 rack: null)  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,712(968) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest24] Revoking previously assigned partitions []  
2018-12-20 13:16:13 [INFO ] 2018-12-20 13:16:13,713(969) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:336): [Consumer clientId=consumer-2, groupId=seekTest24] (Re-)joining group  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,851(4107) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest24] Successfully joined group with generation 1  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,852(4108) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.info(LogContext.java:341): [Consumer clientId=consumer-2, groupId=seekTest24] Setting newly assigned partitions [seek_test_sys-0]  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,881(4137) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 8, value 8.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,884(4140) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 9, value 9.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,885(4141) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 10, value 10.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,886(4142) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 11, value 11.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,886(4142) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 12, value 12.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,886(4142) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 13, value 13.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,887(4143) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 14, value 14.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,887(4143) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 15, value 15.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,887(4143) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:149):  F : offset 16, value 16.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,888(4144) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:155): seek  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,894(4150) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,895(4151) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,896(4152) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,897(4153) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,897(4153) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,897(4153) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,898(4154) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,898(4154) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,898(4154) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,899(4155) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,899(4155) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,899(4155) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,901(4157) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,902(4158) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,904(4160) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,905(4161) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,906(4162) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,906(4162) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,906(4162) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,907(4163) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,907(4163) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,907(4163) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,908(4164) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,908(4164) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,909(4165) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 66, value 66.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,909(4165) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,909(4165) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,910(4166) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,910(4166) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 67, value 67.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,910(4166) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,911(4167) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 68, value 68.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,911(4167) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 69, value 69.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,912(4168) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 70, value 70.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,912(4168) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 66: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,913(4169) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 71, value 71.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,913(4169) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=66, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,914(4170) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 72, value 72.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,915(4171) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 67: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,915(4171) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 73, value 73.  
2018-12-20 13:16:16 [WARN ] 2018-12-20 13:16:16,915(4171) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.warn(LogContext.java:246): [Consumer clientId=consumer-1, groupId=seekTest24] Asynchronous auto-commit of offsets {seek_test_sys-0=OffsetAndMetadata{offset=67, metadata=''}} failed: Commit cannot be completed since the group has already rebalanced and assigned the partitions to another member. This means that the time between subsequent calls to poll() was longer than the configured max.poll.interval.ms, which typically implies that the poll loop is spending too much time message processing. You can address this either by increasing the session timeout or by reducing the maximum size of batches returned in poll() with max.poll.records.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,915(4171) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 68: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [ERROR] 2018-12-20 13:16:16,916(4172) --> [main] org.apache.kafka.common.utils.LogContext$KafkaLogger.error(LogContext.java:301): [Consumer clientId=consumer-1, groupId=seekTest24] Offset commit failed on partition seek_test_sys-0 at offset 69: The coordinator is not aware of this member.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,916(4172) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:160):  E : offset 74, value 74.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,917(4173) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 17, value 17.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,918(4174) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 18, value 18.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,918(4174) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 19, value 19.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,919(4175) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 20, value 20.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,919(4175) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 21, value 21.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,919(4175) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 22, value 22.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,920(4176) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 23, value 23.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,920(4176) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 24, value 24.  
2018-12-20 13:16:16 [INFO ] 2018-12-20 13:16:16,921(4177) --> [main] com.songys.kafka.kafkaSeekTest.assignAndSubscribe(kafkaSeekTest.java:167):  F : offset 25, value 25.  
复制代码

整体验证测试

三个线程，每个线程一个consumer，配置相同：

Thread-1 / 2 ：消费者 assign 相同的topic-partition，进行消费。

Thread-3 ：消费者不进行任何消费，不断拉取kafka中该group.id上一次commit的offset值。 结果：

两个线程Thread-1 和 2中的consumer各自消费了一份全量数据，没有什么问题。

但是Thread-3中的consumer拉到的值部分如下：

2018-12-25 17:25:57 [INFO ] 2018-12-25 17:25:57,967(505) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:57 [INFO ] 2018-12-25 17:25:57,999(537) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,046(584) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,098(636) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,230(768) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,247(785) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,356(894) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,458(996) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,478(1016) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 212
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,532(1070) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 0
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,546(1084) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 181
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,569(1107) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 213
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,583(1121) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 252
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,614(1152) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 261
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,629(1167) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 268
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,638(1176) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 275
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,655(1193) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 283
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,667(1205) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 324
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,677(1215) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 332
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,713(1251) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 356
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,817(1355) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 499
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,854(1392) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 693
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,962(1500) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 850
2018-12-25 17:25:58 [INFO ] 2018-12-25 17:25:58,991(1529) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1242
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,006(1544) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 949
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,020(1558) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 971
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,030(1568) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1006
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,038(1576) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1076
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,076(1614) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1085
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,090(1628) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1282
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,110(1648) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1104
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,134(1672) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1299
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,270(1808) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1303
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,463(2001) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1515
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,600(2138) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 1805
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,692(2230) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 2300
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,803(2341) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 2569
2018-12-25 17:25:59 [INFO ] 2018-12-25 17:25:59,902(2440) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 2923
2018-12-25 17:26:00 [INFO ] 2018-12-25 17:26:00,074(2612) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 2613
2018-12-25 17:26:00 [INFO ] 2018-12-25 17:26:00,097(2635) --> [Thread-2] com.ymxz.kafka.KafkaOffsetAndMeta.run(KafkaOffsetAndMeta.java:35): the offset last committed : offset 2861

复制代码

可以看到两个consumer提交到broker的offset互相影响，但是没有影响到各自的消费，只是kafka broker上保存的offset值不知道到底是哪个consumer提交的。

你可能感兴趣的:(关于kafka中consumer subscribe与asssign的理解)

  • 斤斤计较的婚姻到底有多难？ 白心之岂必有为
    很多人私聊我会问到在哪个人群当中斤斤计较的人最多？我都会回答他，一般婚姻出现问题的斤斤计较的人士会非常多，以我多年经验，在婚姻落的一塌糊涂的人当中，斤斤计较的人数占比在20～30%以上，也就是说10个婚姻出现问题的斤斤计较的人有2-3个有多不减。在婚姻出问题当中，有大量的心理不平衡的、尖酸刻薄的怨妇。在婚姻中仅斤斤计较有两种类型：第一种是物质上的，另一种是精神上的。在物质与精神上抠门已经严重的影响
  • 情绪觉察日记第37天 露露_e800
    今天是家庭关系规划师的第二阶最后一天，慧萍老师帮我做了个案，帮我处理了埋在心底好多年的一份恐惧，并给了我深深的力量！这几天出来学习，爸妈过来婆家帮我带小孩，妈妈出于爱帮我收拾东西，并跟我先生和婆婆产生矛盾，妈妈觉得他们没有照顾好我…。今晚回家见到妈妈，我很欣赏她并赞扬她，妈妈说今晚要跟我睡我说好，当我们俩躺在床上准备睡觉的时候，我握着妈妈的手对她说:妈妈这几天辛苦你了，你看你多利害把我们的家收拾得
  • 芦花鞋一四 许叶晗
    又是在一个寒冷的夏日里，青铜和葵花决定今天一起去卖芦花鞋，奶奶亲手给他们做了一碗热乎乎的粥对他们说:“就靠你们两挣生活费了这碗粥赶紧趁热喝了吧！”于是青铜和葵花喝完了奶奶给她们做的粥，就准备去镇上卖卢花鞋，这回青铜和葵花穿着新的芦花鞋来到了镇上。青铜这回看到了很多人都在卖，用手势表达对葵花说:“这回有好多人在抢我们生意呢！我们必须得吆喝起来。”葵花点了点头。可是谁知他们也大声的叫，卖芦花喽！卖芦花
  • QQ群采集助手，精准引流必备神器 2401_87347160 其他经验分享
    功能概述微信群查找与筛选工具是一款专为微信用户设计的辅助工具，它通过关键词搜索功能，帮助用户快速找到相关的微信群，并提供筛选是否需要验证的群组的功能。主要功能关键词搜索：用户可以输入关键词，工具将自动查找包含该关键词的微信群。筛选功能：工具提供筛选机制，用户可以选择是否只显示需要验证或不需要验证的群组。精准引流：通过上述功能，用户可以更精准地找到目标群组，进行有效的引流操作。3.设备需求该工具可以
  • 关于沟通这件事，项目经理不需要每次都面对面进行 流程大师兄
    很多项目经理都会遇到这样的问题，项目中由于事情太多，根本没有足够的时间去召开会议，那在这种情况下如何去有效地管理项目中的利益相关者？当然，不建议电子邮件也不需要开会的话，建议可以采取下面几种方式来形成有效的沟通，这几种方式可以帮助你努力的通过各种办法来保持和各方面的联系。项目经理首先要问自己几个问题，项目中哪些利益相关者是必须要进行沟通的？可以列出项目中所有的利益相关者清单，同时也整理出项目中哪些
  • 机器学习与深度学习间关系与区别 ℒℴѵℯ心·动ꦿ໊ོ꫞ 人工智能学习深度学习python
    一、机器学习概述定义机器学习（MachineLearning,ML）是一种通过数据驱动的方法，利用统计学和计算算法来训练模型，使计算机能够从数据中学习并自动进行预测或决策。机器学习通过分析大量数据样本，识别其中的模式和规律，从而对新的数据进行判断。其核心在于通过训练过程，让模型不断优化和提升其预测准确性。主要类型1.监督学习（SupervisedLearning）监督学习是指在训练数据集中包含输入
  • android系统selinux中添加新属性property 辉色投像
    1.定位/android/system/sepolicy/private/property_contexts声明属性开头：persist.charge声明属性类型：u:object_r:system_prop:s0图12.定位到android/system/sepolicy/public/domain.te删除neverallow{domain-init}default_prop:property
  • 铭刻于星（四十二） 随风至
    69夜晚，绍敏同学做完功课后，看了眼房外，没听到动静才敢从书包的夹层里拿出那个心形纸团。折痕压得很深，都有些旧了，想来是已经写好很久了。绍敏同学慢慢地、轻轻地捏开折叠处，待到全部拆开后，又反复抚平纸张，然后仔细地一字字默看。只是开头的三个字是第一次看到，让她心漏跳了几拍。“亲爱的绍敏：从四年级的时候，我就喜欢你了，但是我一直不敢说，怕影响你学习。六年级的时候听说有人跟你表白，你接受了，我很难过，但
  • 底层逆袭到底有多难，不甘平凡的你准备好了吗？让吴起给你说说 造命者说
    底层逆袭到底有多难，不甘平凡的你准备好了吗？让吴起给你说说我叫吴起，生于公元前440年的战国初期，正是群雄并起、天下纷争不断的时候。后人说我是军事家、政治家、改革家，是兵家代表人物。评价我一生历仕鲁、魏、楚三国，通晓兵家、法家、儒家三家思想，在内政军事上都有极高的成就。周安王二十一年（公元前381年），因变法得罪守旧贵族，被人乱箭射死。我出生在卫国一个“家累万金”的富有家庭，从年轻时候起就不甘平凡
  • 2020-01-25 晴岚85
    郑海燕坚持分享590天2020.1.24在生活中只存在两个问题。一个问题是：你知道想要达成的目标是什么，但却不知道如何才能达成；另一个问题是：你不知道你的目标是什么。前一个是行动的问题，后一个是结果的问题。通过制定具体的下一步行动，可以解决不知道如何开始行动的问题。而通过去想象结果，对结果做预估，可以解决找不着目标的问题。对于所有吸引我们注意力，想要完成的任务，你可以先想象一下，预期的结果究竟是什
  • 随笔 | 仙一般的灵气 海思沧海
    仙岛今天，我看了你全部，似乎已经进入你的世界我不知道，这是否是梦幻，还是你仙一般的灵气吸引了我也许每一个人都要有一份属于自己的追求，这样才能够符合人生的梦想，生活才能够充满着阳光与快乐我不知道，我为什么会这样的感叹，是在感叹自己的人生，还是感叹自己一直没有孜孜不倦的追求只感觉虚度了光阴，每天活在自己的梦中，活在一个不真实的世界是在逃避自己，还是在逃避周围的一切有时候我嘲笑自己，嘲笑自己如此的虚无，
  • 想家 爆米花机
    也许不同于大家对家乡的思念，我对家乡甚至是疯狂的不舍。还未踏出车站就感觉到幸福，我享受这里的夕阳、这里的浓烈柴火味、这里每一口家常菜。我是宅女，我贪恋家的安逸。刚刚踏出大学校门，初出茅庐，无法适应每年只能国庆和春节回家。我焦虑、失眠、无端发脾气，是无法适应工作的节奏，是无法接受我将一步步离开家乡的事实。我不想承认自己胸无大志，选择再次踏上征程。图片发自App
  • 【iOS】MVC设计模式 Magnetic_h iosmvc设计模式objective-c学习ui
    MVC前言如何设计一个程序的结构，这是一门专门的学问，叫做"架构模式"（architecturalpattern），属于编程的方法论。MVC模式就是架构模式的一种。它是Apple官方推荐的App开发架构，也是一般开发者最先遇到、最经典的架构。MVC各层controller层Controller/ViewController/VC（控制器）负责协调Model和View，处理大部分逻辑它将数据从Mod
  • OC语言多界面传值五大方式 Magnetic_h iosui学习objective-c开发语言
    前言在完成暑假仿写项目时，遇到了许多需要用到多界面传值的地方，这篇博客来总结一下比较常用的五种多界面传值的方式。属性传值属性传值一般用前一个界面向后一个界面传值，简单地说就是通过访问后一个视图控制器的属性来为它赋值，通过这个属性来做到从前一个界面向后一个界面传值。首先在后一个界面中定义属性@interfaceBViewController:UIViewController@propertyNSSt
  • 一百九十四章. 自相矛盾 巨木擎天
    唉！就这么一夜，林子感觉就像过了很多天似的，先是回了阳间家里，遇到了那么多不可思议的事情儿。特别是小伙伴们，第二次与自己见面时，僵硬的表情和恐怖的气氛，让自己如坐针毡，打从心眼里难受！还有东子，他现在还好吗？有没有被人欺负？护城河里的小鱼小虾们，还都在吗？水不会真的干枯了吧？那对相亲相爱漂亮的太平鸟儿，还好吧！春天了，到了做窝、下蛋、喂养小鸟宝宝的时候了，希望它们都能够平安啊！虽然没有看见家人，也
  • UI学习——cell的复用和自定义cell Magnetic_h ui学习
    目录cell的复用手动（非注册）自动（注册）自定义cellcell的复用在iOS开发中，单元格复用是一种提高表格（UITableView）和集合视图（UICollectionView）滚动性能的技术。当一个UITableViewCell或UICollectionViewCell首次需要显示时，如果没有可复用的单元格，则视图会创建一个新的单元格。一旦这个单元格滚动出屏幕，它就不会被销毁。相反，它被添
  • element实现动态路由+面包屑 软件技术NINI vue案例vue.js前端
    el-breadcrumb是ElementUI组件库中的一个面包屑导航组件，它用于显示当前页面的路径，帮助用户快速理解和导航到应用的各个部分。在Vue.js项目中，如果你已经安装了ElementUI，就可以很方便地使用el-breadcrumb组件。以下是一个基本的使用示例：安装ElementUI（如果你还没有安装的话）:你可以通过npm或yarn来安装ElementUI。bash复制代码npmi
  • 10月|愿你的青春不负梦想-读书笔记-01 Tracy的小书斋
    本书的作者是俞敏洪，大家都很熟悉他了吧。俞敏洪老师是我行业的领头羊吧，也是我事业上的偶像。本日摘录他书中第一章中的金句：『一个人如果什么目标都没有，就会浑浑噩噩，感觉生命中缺少能量。能给我们能量的，是对未来的期待。第一件事，我始终为了进步而努力。与其追寻全世界的骏马，不如种植丰美的草原，到时骏马自然会来。第二件事，我始终有阶段性的目标。什么东西能给我能量？答案是对未来的期待。』读到这里的时候，我便
  • C语言宏函数 南林yan C语言c语言
    一、什么是宏函数？通过宏定义的函数是宏函数。如下，编译器在预处理阶段会将Add(x,y)替换为((x)*(y))#defineAdd(x,y)((x)*(y))#defineAdd(x,y)((x)*(y))intmain(){inta=10;intb=20;intd=10;intc=Add(a+d,b)*2;cout<
  • 地推话术，如何应对地推过程中家长的拒绝 校师学
    相信校长们在做地推的时候经常遇到这种情况：市场专员反馈家长不接单，咨询师反馈难以邀约这些家长上门，校区地推疲软，招生难。为什么？仅从地推层面分析，一方面因为家长受到的信息轰炸越来越多，对信息越来越“免疫”；而另一方面地推人员的专业能力和营销话术没有提高，无法应对家长的拒绝，对有意向的家长也不知如何跟进，眼睁睁看着家长走远；对于家长的疑问，更不知道如何有技巧地回答，机会白白流失。由于回答没技巧和专业
  • 谢谢你们，爱你们！ 鹿游儿
    昨天家人去泡温泉，二个孩子也带着去，出发前一晚，匆匆下班，赶回家和孩子一起收拾。饭后，我拿出笔和本子（上次去澳门时做手帐的本子）写下了1\2\3\4\5\6\7\8\9,让后让小壹去思考，带什么出发去旅游呢？她在对应的数字旁边画上了，泳衣、泳圈、肖恩、内衣内裤、tapuy、拖鞋……画完后，就让她自己对着这个本子，将要带的，一一带上，没想到这次带的书还是这本《便便工厂》(晚上姑婆发照片过来，妹妹累得
  • C语言如何定义宏函数？ 小九格物 c语言
    在C语言中，宏函数是通过预处理器定义的，它在编译之前替换代码中的宏调用。宏函数可以模拟函数的行为，但它们不是真正的函数，因为它们在编译时不会进行类型检查，也不会分配存储空间。宏函数的定义通常使用#define指令，后面跟着宏的名称和参数列表，以及宏展开后的代码。宏函数的定义方式：1.基本宏函数：这是最简单的宏函数形式，它直接定义一个表达式。#defineSQUARE(x)((x)*(x))2.带参
  • 微服务下功能权限与数据权限的设计与实现 nbsaas-boot 微服务java架构
    在微服务架构下，系统的功能权限和数据权限控制显得尤为重要。随着系统规模的扩大和微服务数量的增加，如何保证不同用户和服务之间的访问权限准确、细粒度地控制，成为设计安全策略的关键。本文将讨论如何在微服务体系中设计和实现功能权限与数据权限控制。1.功能权限与数据权限的定义功能权限：指用户或系统角色对特定功能的访问权限。通常是某个用户角色能否执行某个操作，比如查看订单、创建订单、修改用户资料等。数据权限：
  • 理解Gunicorn：Python WSGI服务器的基石 范范0825 ipythonlinux运维
    理解Gunicorn：PythonWSGI服务器的基石介绍Gunicorn，全称GreenUnicorn，是一个为PythonWSGI（WebServerGatewayInterface）应用设计的高效、轻量级HTTP服务器。作为PythonWeb应用部署的常用工具，Gunicorn以其高性能和易用性著称。本文将介绍Gunicorn的基本概念、安装和配置，帮助初学者快速上手。1.什么是Gunico
  • 小丽成长记（四十三） 玲玲54321
    小丽发现，即使她好不容易调整好自己的心态下一秒总会有不确定的伤脑筋的事出现，一个接一个的问题，人生就没有停下的时候，小问题不断出现。不过她今天看的书，她接受了人生就是不确定的，厉害的人就是不断创造确定性，在Ta的领域比别人多的确定性就能让自己脱颖而出，显示价值从而获得的比别人多的利益。正是这样的原因，因为从前修炼自己太少，使得她现在在人生道路上打怪起来困难重重，她似乎永远摆脱不了那种无力感，有种习
  • 学点心理知识，呵护孩子健康 静候花开_7090
    昨天听了华中师范大学教育管理学系副教授张玲老师的《哪里才是学生心理健康的最后庇护所，超越教育与技术的思考》的讲座。今天又重新学习了一遍，收获匪浅。张玲博士也注意到了当今社会上的孩子由于心理问题导致的自残、自杀及伤害他人等恶性事件。她向我们普及了一个重要的命题，她说心理健康的一些基本命题，我们与我们通常的一些教育命题是不同的，她还举了几个例子，让我们明白我们原来以为的健康并非心理学上的健康。比如如果
  • 2021年12月19日，春蕾教育集团团建活动感受——黄晓丹 黄错错加油
    感受:1.从陌生到熟悉的过程。游戏环节让我们在轻松的氛围中得到了锻炼，也增长了不少知识。2.游戏过程中，我们贡献的是个人力量，展现的是团队的力量。它磨合的往往不止是工作的熟悉，更是观念上契合度的贴近。3.这和工作是一样的道理。在各自的岗位上，每个人摆正自己的位置、各司其职充分发挥才能，并团结一致劲往一处使，才能实现最大的成功。新知:1.团队精神需要不断地创新。过去，人们把创新看作是冒风险，现在人们
  • Cell Insight | 单细胞测序技术又一新发现，可用于HIV-1和Mtb共感染个体诊断 尐尐呅
    结核病是艾滋病合并其他疾病中导致患者死亡的主要原因。其中结核病由结核分枝杆菌（Mycobacteriumtuberculosis,Mtb）感染引起，获得性免疫缺陷综合症（艾滋病）由人免疫缺陷病毒（Humanimmunodeficiencyvirustype1,HIV-1）感染引起。国家感染性疾病临床医学研究中心/深圳市第三人民医院张国良团队携手深圳华大生命科学研究院吴靓团队，共同研究得出单细胞测序
  • c++ 的iostream 和 c++的stdio的区别和联系 黄卷青灯77 c++算法开发语言iostreamstdio
    在C++中，iostream和C语言的stdio.h都是用于处理输入输出的库，但它们在设计、用法和功能上有许多不同。以下是两者的区别和联系：区别1.编程风格iostream（C++风格）：C++标准库中的输入输出流类库，支持面向对象的输入输出操作。典型用法是cin（输入）和cout（输出），使用>操作符来处理数据。更加类型安全，支持用户自定义类型的输入输出。#includeintmain(){in
  • 瑶池防线 谜影梦蝶
    冥华虽然逃过了影梦的军队，但他是一个忠臣，他选择上报战况。败给影梦后成逃兵，高层亡尔还活着，七重天失守......随便一条，即可处死冥华。冥华自然是知道以仙界高层的习性此信一发自己必死无疑，但他还选择上报实情，因为责任。同样此信送到仙宫后，知道此事的人，大多数人都认定冥华要完了，所以上到仙界高层，下到扫大街的，包括冥华自己，全都准备好迎接冥华之死。如果仙界现在还属于两方之争的话，冥华必死无疑。然而
  • Spring中@Value注解，需要注意的地方 无量 springbean@Valuexml
    Spring 3以后,支持@Value注解的方式获取properties文件中的配置值，简化了读取配置文件的复杂操作 1、在applicationContext.xml文件(或引用文件中)中配置properties文件 <bean id="appProperty" class="org.springframework.beans.fac
  • mongoDB 分片 开窍的石头 mongodb
        mongoDB的分片。要mongos查询数据时候 先查询configsvr看数据在那台shard上，configsvr上边放的是metar信息，指的是那条数据在那个片上。由此可以看出mongo在做分片的时候咱们至少要有一个configsvr,和两个以上的shard（片）信息。     第一步启动两台以上的mongo服务 &nb
  • OVER(PARTITION BY)函数用法 0624chenhong oracle
    这篇写得很好，引自 http://www.cnblogs.com/lanzi/archive/2010/10/26/1861338.html OVER(PARTITION BY)函数用法 2010年10月26日 OVER(PARTITION BY)函数介绍 开窗函数        &nb
  • Android开发中，ADB server didn't ACK 解决方法 一炮送你回车库 Android开发
    首先通知：凡是安装360、豌豆荚、腾讯管家的全部卸载，然后再尝试。   一直没搞明白这个问题咋出现的，但今天看到一个方法，搞定了！原来是豌豆荚占用了 5037 端口导致。 参见原文章：一个豌豆荚引发的血案——关于ADB server didn't ACK的问题 简单来讲，首先将Windows任务进程中的豌豆荚干掉，如果还是不行，再继续按下列步骤排查。 &nb
  • canvas中的像素绘制问题 换个号韩国红果果 JavaScriptcanvas
    pixl的绘制，1.如果绘制点正处于相邻像素交叉线，绘制x像素的线宽，则从交叉线分别向前向后绘制x/2个像素，如果x/2是整数，则刚好填满x个像素，如果是小数，则先把整数格填满，再去绘制剩下的小数部分，绘制时，是将小数部分的颜色用来除以一个像素的宽度，颜色会变淡。所以要用整数坐标来画的话（即绘制点正处于相邻像素交叉线时），线宽必须是2的整数倍。否则会出现不饱满的像素。 2.如果绘制点为一个像素的
  • 编码乱码问题 灵静志远 javajvmjsp编码
    1、JVM中单个字符占用的字节长度跟编码方式有关，而默认编码方式又跟平台是一一对应的或说平台决定了默认字符编码方式；2、对于单个字符：ISO-8859-1单字节编码，GBK双字节编码，UTF-8三字节编码；因此中文平台(中文平台默认字符集编码GBK)下一个中文字符占2个字节，而英文平台(英文平台默认字符集编码Cp1252(类似于ISO-8859-1))。 3、getBytes()、getByte
  • java 求几个月后的日期 darkranger calendargetinstance
    Date plandate = planDate.toDate(); SimpleDateFormat df = new SimpleDateFormat("yyyy-MM-dd"); Calendar cal = Calendar.getInstance(); cal.setTime(plandate); // 取得三个月后时间 cal.add(Calendar.M
  • 数据库设计的三大范式（通俗易懂） aijuans 数据库复习
    关系数据库中的关系必须满足一定的要求。满足不同程度要求的为不同范式。数据库的设计范式是数据库设计所需要满足的规范。只有理解数据库的设计范式，才能设计出高效率、优雅的数据库，否则可能会设计出错误的数据库. 目前，主要有六种范式：第一范式、第二范式、第三范式、BC范式、第四范式和第五范式。满足最低要求的叫第一范式，简称1NF。在第一范式基础上进一步满足一些要求的为第二范式，简称2NF。其余依此类推。
  • 想学工作流怎么入手 atongyeye jbpm
    工作流在工作中变得越来越重要，很多朋友想学工作流却不知如何入手。 很多朋友习惯性的这看一点，那了解一点，既不系统，也容易半途而废。好比学武功，最好的办法是有一本武功秘籍。研究明白，则犹如打通任督二脉。 系统学习工作流，很重要的一本书《JBPM工作流开发指南》。 本人苦苦学习两个月，基本上可以解决大部分流程问题。整理一下学习思路，有兴趣的朋友可以参考下。 1  首先要
  • Context和SQLiteOpenHelper创建数据库 百合不是茶 androidContext创建数据库
           一直以为安卓数据库的创建就是使用SQLiteOpenHelper创建,但是最近在android的一本书上看到了Context也可以创建数据库,下面我们一起分析这两种方式创建数据库的方式和区别,重点在SQLiteOpenHelper     一:SQLiteOpenHelper创建数据库:   1,SQLi
  • 浅谈group by和distinct bijian1013 oracle数据库group bydistinct
            group by和distinct只了去重意义一样，但是group by应用范围更广泛些，如分组汇总或者从聚合函数里筛选数据等。         譬如：统计每id数并且只显示数大于3 select id ,count(id) from ta
  • vi opertion 征客丶 macoprationvi
    进入 command mode （命令行模式） 按 esc 键 再按 shift + 冒号 注：以下命令中 带 $ 【在命令行模式下进行】，不带 $ 【在非命令行模式下进行】 一、文件操作 1.1、强制退出不保存 $ q! 1.2、保存 $ w 1.3、保存并退出 $ wq 1.4、刷新或重新加载已打开的文件 $ e 二、光标移动 2.1、跳到指定行 数字
  • 【Spark十四】深入Spark RDD第三部分RDD基本API bit1129 spark
      对于K/V类型的RDD,如下操作是什么含义？   val rdd = sc.parallelize(List(("A",3),("C",6),("A",1),("B",5)) rdd.reduceByKey(_+_).collect  reduceByKey在这里的操作，是把
  • java类加载机制 BlueSkator java虚拟机
    java类加载机制 1.java类加载器的树状结构 引导类加载器 ^ | 扩展类加载器 ^ | 系统类加载器 java使用代理模式来完成类加载，java的类加载器也有类似于继承的关系，引导类是最顶层的加载器，它是所有类的根加载器，它负责加载java核心库。当一个类加载器接到装载类到虚拟机的请求时，通常会代理给父类加载器，若已经是根加载器了，就自己完成加载。 虚拟机区分一个Cla
  • 动态添加文本框 BreakingBad 文本框
      <script>     var num=1; function AddInput() {      var str="";     str+="<input 
  • 读《研磨设计模式》-代码笔记-单例模式 bylijinnan java设计模式
    声明： 本文只为方便我个人查阅和理解，详细的分析以及源代码请移步 原作者的博客http://chjavach.iteye.com/ public class Singleton { } /* * 懒汉模式。注意，getInstance如果在多线程环境中调用，需要加上synchronized，否则存在线程不安全问题 */ class LazySingleton
  • iOS应用打包发布常见问题 chenhbc iosiOS发布iOS上传iOS打包
    这个月公司安排我一个人做iOS客户端开发，由于急着用，我先发布一个版本，由于第一次发布iOS应用，期间出了不少问题，记录于此。   1、使用Application Loader 发布时报错：Communication error.please use diagnostic mode to check connectivity.you need to have outbound acc
  • 工作流复杂拓扑结构处理新思路 comsci 设计模式工作算法企业应用OO
    我们走的设计路线和国外的产品不太一样，不一样在哪里呢？  国外的流程的设计思路是通过事先定义一整套规则(类似XPDL)来约束和控制流程图的复杂度(我对国外的产品了解不够多，仅仅是在有限的了解程度上面提出这样的看法)，从而避免在流程引擎中处理这些复杂的图的问题，而我们却没有通过事先定义这样的复杂的规则来约束和降低用户自定义流程图的灵活性，这样一来，在引擎和流程流转控制这一个层面就会遇到很
  • oracle 11g新特性Flashback data archive daizj oracle
    1. 什么是flashback data archive Flashback data archive是oracle 11g中引入的一个新特性。Flashback archive是一个新的数据库对象，用于存储一个或多表的历史数据。Flashback archive是一个逻辑对象，概念上类似于表空间。实际上flashback archive可以看作是存储一个或多个表的所有事务变化的逻辑空间。
  • 多叉树:2-3-4树 dieslrae 树
        平衡树多叉树,每个节点最多有4个子节点和3个数据项,2,3,4的含义是指一个节点可能含有的子节点的个数,效率比红黑树稍差.一般不允许出现重复关键字值.2-3-4树有以下特征:     1、有一个数据项的节点总是有2个子节点(称为2-节点)     2、有两个数据项的节点总是有3个子节点(称为3-节
  • C语言学习七动态分配 malloc的使用 dcj3sjt126com clanguagemalloc
    /* 2013年3月15日15:16:24 malloc 就memory(内存) allocate(分配)的缩写 本程序没有实际含义，只是理解使用 */ # include <stdio.h> # include <malloc.h> int main(void) { int i = 5; //分配了4个字节 静态分配 int * p
  • Objective-C编码规范[译] dcj3sjt126com 代码规范
      原文链接 : The official raywenderlich.com Objective-C style guide 原文作者 : raywenderlich.com Team 译文出自 : raywenderlich.com Objective-C编码规范 译者 : Sam Lau
  • 0.性能优化-目录 frank1234 性能优化
    从今天开始笔者陆续发表一些性能测试相关的文章，主要是对自己前段时间学习的总结，由于水平有限，性能测试领域很深，本人理解的也比较浅，欢迎各位大咖批评指正。 主要内容包括： 一、性能测试指标 吞吐量、TPS、响应时间、负载、可扩展性、PV、思考时间 http://frank1234.iteye.com/blog/2180305 二、性能测试策略 生产环境相同 基准测试 预热等 htt
  • Java父类取得子类传递的泛型参数Class类型 happyqing java泛型父类子类Class
      import java.lang.reflect.ParameterizedType; import java.lang.reflect.Type; import org.junit.Test; abstract class BaseDao<T> { public void getType() { //Class<E> clazz =
  • 跟我学SpringMVC目录汇总贴、PDF下载、源码下载 jinnianshilongnian springMVC
      ----广告-------------------------------------------------------------- 网站核心商详页开发 掌握Java技术，掌握并发/异步工具使用，熟悉spring、ibatis框架； 掌握数据库技术，表设计和索引优化，分库分表/读写分离； 了解缓存技术，熟练使用如Redis/Memcached等主流技术； 了解Ngin
  • the HTTP rewrite module requires the PCRE library 流浪鱼 rewrite
    ./configure: error: the HTTP rewrite module requires the PCRE library. 模块依赖性Nginx需要依赖下面3个包 1. gzip 模块需要 zlib 库 ( 下载: http://www.zlib.net/ ) 2. rewrite 模块需要 pcre 库 ( 下载: http://www.pcre.org/ ) 3. s
  • 第12章 Ajax（中） onestopweb Ajax
    index.html <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html xmlns="http://www.w3.org/
  • Optimize query with Query Stripping in Web Intelligence blueoxygen BO
    http://wiki.sdn.sap.com/wiki/display/BOBJ/Optimize+query+with+Query+Stripping+in+Web+Intelligence and a very straightfoward video http://www.sdn.sap.com/irj/scn/events?rid=/library/uuid/40ec3a0c-936
  • Java开发者写SQL时常犯的10个错误 tomcat_oracle javasql
    1、不用PreparedStatements 　　有意思的是，在JDBC出现了许多年后的今天，这个错误依然出现在博客、论坛和邮件列表中，即便要记住和理解它是一件很简单的事。开发者不使用PreparedStatements的原因可能有如下几个： 　　他们对PreparedStatements不了解 　　他们认为使用PreparedStatements太慢了 　　他们认为写Prepar
  • 世纪互联与结盟有感 阿尔萨斯
    10月10日，世纪互联与（Foxcon）签约成立合资公司，有感。 全球电子制造业巨头（全球500强企业）与世纪互联共同看好IDC、云计算等业务在中国的增长空间，双方迅速果断出手，在资本层面上达成合作，此举体现了全球电子制造业巨头对世纪互联IDC业务的欣赏与信任，另一方面反映出世纪互联目前良好的运营状况与广阔的发展前景。 众所周知，精于电子产品制造（世界第一），对于世纪互联而言，能够与结盟
按字母分类： ABCDEFGHIJKLMNOPQRSTUVWXYZ其他