Question

当系统处于空闲状态并且启用了自动提交偏移量时，有时会收到非常奇怪的错误消息： 2019-02-15 16:31:42,742 ERROR[pool-9-thread-1] org.apache.kafka.clients.consumer.internals.ConsumerCoordinator - [Consumer clientId=ClientVerticle_Thread[pool-9-thread-1,5,main], groupId=mygroupid] Offset commit failed on partition MYYOPIC-5 at offset 0: The request timed out.', u'[Consumer clientId=ClinetVerticle_Thread[pool-9-thread-1,5,main], groupId=mygroupid] Offset commit failed on partition MYTOPIC-5 at offset 0: The request timed out 有关设置的一些数据会重现此错误：

启用了自动提交
仅当offset为0（没有消息访问分区）时，才会发生此错误
仅当主题附加了多个消费者组（在这种情况下为2个不同的消费者组）时，才会发生此错误
主题分区数为：6
（每个消费者组）消费者线程数为：3
消费者设置：


    config.put("client.id", InetAddress.getLocalHost().getHostAddress() + "_" + clientId + "_" + Thread.currentThread());
                    config.put("group.id", consumerGroupName);
                    config.put("bootstrap.servers", clusterUrl);
                    config.put("auto.offset.reset", "latest");
                    config.put("heartbeat.interval.ms", 3000);
                    config.put("session.timeout.ms", 10000);
                    config.put("request.timeout.ms", 10000 + 3000);
                    config.put("enable.auto.commit", true);
                    config.put("auto.commit.interval.ms", 2000);
                    config.put("key.deserializer", StringDeserializer.class.getName());
                    config.put("value.deserializer", StringDeserializer.class.getName());
                    config.put("fetch.min.bytes", 1);
                    config.put("max.partition.fetch.bytes", 1048576);
                    config.put("fetch.max.wait.ms", 500);

谢谢

Kafka-在分区0的MYTOPIC-X上，偏移提交失败

0 个答案: