1.概述
最近有同學諮詢Kafka的消費和心跳機制,今天筆者將通過這篇部落格來逐一介紹這些內容。
2.內容
2.1 Kafka消費
首先,我們來看看消費。Kafka提供了非常簡單的消費API,使用者只需初始化Kafka的Broker Server地址,然後例項化KafkaConsumer類即可拿到Topic中的資料。一個簡單的Kafka消費例項程式碼如下所示:
public class JConsumerSubscribe extends Thread { public static void main(String[] args) { JConsumerSubscribe jconsumer = new JConsumerSubscribe(); jconsumer.start(); } /** 初始化Kafka叢集資訊. */ private Properties configure() { Properties props = new Properties(); props.put("bootstrap.servers", "dn1:9092,dn2:9092,dn3:9092");// 指定Kafka叢集地址 props.put("group.id", "ke");// 指定消費者組 props.put("enable.auto.commit", "true");// 開啟自動提交 props.put("auto.commit.interval.ms", "1000");// 自動提交的時間間隔 // 反序列化訊息主鍵 props.put("key.deserializer", "org.apache.kafka.common.serialization.StringDeserializer"); // 反序列化消費記錄 props.put("value.deserializer", "org.apache.kafka.common.serialization.StringDeserializer"); return props; } /** 實現一個單執行緒消費者. */ @Override public void run() { // 建立一個消費者例項物件 KafkaConsumer<String, String> consumer = new KafkaConsumer<>(configure()); // 訂閱消費主題集合 consumer.subscribe(Arrays.asList("test_kafka_topic")); // 實時消費標識 boolean flag = true; while (flag) { // 獲取主題訊息資料 ConsumerRecords<String, String> records = consumer.poll(Duration.ofMillis(100)); for (ConsumerRecord<String, String> record : records) // 迴圈列印訊息記錄 System.out.printf("offset = %d, key = %s, value = %s%n", record.offset(), record.key(), record.value()); } // 出現異常關閉消費者物件 consumer.close(); } }
上述程式碼我們就可以非常便捷的拿到Topic中的資料。但是,當我們呼叫poll方法拉取資料的時候,Kafka Broker Server做了那些事情。接下來,我們可以去看看原始碼的實現細節。核心程式碼如下:
org.apache.kafka.clients.consumer.KafkaConsumer
private ConsumerRecords<K, V> poll(final long timeoutMs, final boolean includeMetadataInTimeout) { acquireAndEnsureOpen(); try { if (timeoutMs < 0) throw new IllegalArgumentException("Timeout must not be negative"); if (this.subscriptions.hasNoSubscriptionOrUserAssignment()) { throw new IllegalStateException("Consumer is not subscribed to any topics or assigned any partitions"); } // poll for new data until the timeout expires long elapsedTime = 0L; do { client.maybeTriggerWakeup(); final long metadataEnd; if (includeMetadataInTimeout) { final long metadataStart = time.milliseconds(); if (!updateAssignmentMetadataIfNeeded(remainingTimeAtLeastZero(timeoutMs, elapsedTime))) { return ConsumerRecords.empty(); } metadataEnd = time.milliseconds(); elapsedTime += metadataEnd - metadataStart; } else { while (!updateAssignmentMetadataIfNeeded(Long.MAX_VALUE)) { log.warn("Still waiting for metadata"); } metadataEnd = time.milliseconds(); } final Map<TopicPartition, List<ConsumerRecord<K, V>>> records = pollForFetches(remainingTimeAtLeastZero(timeoutMs, elapsedTime)); if (!records.isEmpty()) { // before returning the fetched records, we can send off the next round of fetches // and avoid block waiting for their responses to enable pipelining while the user // is handling the fetched records. // // NOTE: since the consumed position has already been updated, we must not allow // wakeups or any other errors to be triggered prior to returning the fetched records. if (fetcher.sendFetches() > 0 || client.hasPendingRequests()) { client.pollNoWakeup(); } return this.interceptors.onConsume(new ConsumerRecords<>(records)); } final long fetchEnd = time.milliseconds(); elapsedTime += fetchEnd - metadataEnd; } while (elapsedTime < timeoutMs); return ConsumerRecords.empty(); } finally { release(); } }
上述程式碼中有個方法pollForFetches,它的實現邏輯如下:
private Map<TopicPartition, List<ConsumerRecord<K, V>>> pollForFetches(final long timeoutMs) { final long startMs = time.milliseconds(); long pollTimeout = Math.min(coordinator.timeToNextPoll(startMs), timeoutMs); // if data is available already, return it immediately final Map<TopicPartition, List<ConsumerRecord<K, V>>> records = fetcher.fetchedRecords(); if (!records.isEmpty()) { return records; } // send any new fetches (won't resend pending fetches) fetcher.sendFetches(); // We do not want to be stuck blocking in poll if we are missing some positions // since the offset lookup may be backing off after a failure // NOTE: the use of cachedSubscriptionHashAllFetchPositions means we MUST call // updateAssignmentMetadataIfNeeded before this method. if (!cachedSubscriptionHashAllFetchPositions && pollTimeout > retryBackoffMs) { pollTimeout = retryBackoffMs; } client.poll(pollTimeout, startMs, () -> { // since a fetch might be completed by the background thread, we need this poll condition // to ensure that we do not block unnecessarily in poll() return !fetcher.hasCompletedFetches(); }); // after the long poll, we should check whether the group needs to rebalance // prior to returning data so that the group can stabilize faster if (coordinator.rejoinNeededOrPending()) { return Collections.emptyMap(); } return fetcher.fetchedRecords(); }
上述程式碼中加粗的位置,我們可以看出每次消費者客戶端拉取資料時,通過poll方法,先呼叫fetcher中的fetchedRecords函式,如果獲取不到資料,就會發起一個新的sendFetches請求。而在消費資料的時候,每個批次從Kafka Broker Server中拉取資料是有最大資料量限制,預設是500條,由屬性(max.poll.records)控制,可以在客戶端中設定該屬性值來調整我們消費時每次拉取資料的量。
提示:
這裡需要注意的是,max.poll.records返回的是一個poll請求的資料總和,與多少個分割槽無關。因此,每次消費從所有分割槽中拉取Topic的資料的總條數不會超過max.poll.records所設定的值。
而在Fetcher的類中,在sendFetches方法中有限制拉取資料容量的限制,由屬性(max.partition.fetch.bytes),預設1MB。可能會有這樣一個場景,當滿足max.partition.fetch.bytes限制條件,如果需要Fetch出10000條記錄,每次預設500條,那麼我們需要執行20次才能將這一次通過網路發起的請求全部Fetch完畢。
這裡,可能有同學有疑問,我們不能將預設的max.poll.records屬性值調到10000嗎?可以調,但是還有個屬性需要一起配合才可以,這個就是每次poll的超時時間(Duration.ofMillis(100)),這裡需要根據你的實際每條資料的容量大小來確定設定超時時間,如果你將最大值調到10000,當你每條記錄的容量很大時,超時時間還是100ms,那麼可能拉取的資料少於10000條。
而這裡,還有另外一個需要注意的事情,就是會話超時的問題。session.timeout.ms預設是10s,group.min.session.timeout.ms預設是6s,group.max.session.timeout.ms預設是30min。當你在處理消費的業務邏輯的時候,如果在10s內沒有處理完,那麼消費者客戶端就會與Kafka Broker Server斷開,消費掉的資料,產生的offset就沒法提交給Kafka,因為Kafka Broker Server此時認為該消費者程式已經斷開,而即使你設定了自動提交屬性,或者設定auto.offset.reset屬性,你消費的時候還是會出現重複消費的情況,這就是因為session.timeout.ms超時的原因導致的。
2.2 心跳機制
上面在末尾的時候,說到會話超時的情況導致訊息重複消費,為什麼會有超時?有同學會有這樣的疑問,我的消費者執行緒明明是啟動的,也沒有退出,為啥消費不到Kafka的訊息呢?消費者組也查不到我的ConsumerGroupID呢?這就有可能是超時導致的,而Kafka是通過心跳機制來控制超時,心跳機制對於消費者客戶端來說是無感的,它是一個非同步執行緒,當我們啟動一個消費者例項時,心跳執行緒就開始工作了。
在org.apache.kafka.clients.consumer.internals.AbstractCoordinator中會啟動一個HeartbeatThread執行緒來定時傳送心跳和檢測消費者的狀態。每個消費者都有個org.apache.kafka.clients.consumer.internals.ConsumerCoordinator,而每個ConsumerCoordinator都會啟動一個HeartbeatThread執行緒來維護心跳,心跳資訊存放在org.apache.kafka.clients.consumer.internals.Heartbeat中,宣告的Schema如下所示:
private final int sessionTimeoutMs; private final int heartbeatIntervalMs; private final int maxPollIntervalMs; private final long retryBackoffMs; private volatile long lastHeartbeatSend; private long lastHeartbeatReceive; private long lastSessionReset; private long lastPoll; private boolean heartbeatFailed;
心跳執行緒中的run方法實現程式碼如下:
public void run() { try { log.debug("Heartbeat thread started"); while (true) { synchronized (AbstractCoordinator.this) { if (closed) return; if (!enabled) { AbstractCoordinator.this.wait(); continue; } if (state != MemberState.STABLE) { // the group is not stable (perhaps because we left the group or because the coordinator // kicked us out), so disable heartbeats and wait for the main thread to rejoin. disable(); continue; } client.pollNoWakeup(); long now = time.milliseconds(); if (coordinatorUnknown()) { if (findCoordinatorFuture != null || lookupCoordinator().failed()) // the immediate future check ensures that we backoff properly in the case that no // brokers are available to connect to. AbstractCoordinator.this.wait(retryBackoffMs); } else if (heartbeat.sessionTimeoutExpired(now)) { // the session timeout has expired without seeing a successful heartbeat, so we should // probably make sure the coordinator is still healthy. markCoordinatorUnknown(); } else if (heartbeat.pollTimeoutExpired(now)) { // the poll timeout has expired, which means that the foreground thread has stalled // in between calls to poll(), so we explicitly leave the group. maybeLeaveGroup(); } else if (!heartbeat.shouldHeartbeat(now)) { // poll again after waiting for the retry backoff in case the heartbeat failed or the // coordinator disconnected AbstractCoordinator.this.wait(retryBackoffMs); } else { heartbeat.sentHeartbeat(now); sendHeartbeatRequest().addListener(new RequestFutureListener<Void>() { @Override public void onSuccess(Void value) { synchronized (AbstractCoordinator.this) { heartbeat.receiveHeartbeat(time.milliseconds()); } } @Override public void onFailure(RuntimeException e) { synchronized (AbstractCoordinator.this) { if (e instanceof RebalanceInProgressException) { // it is valid to continue heartbeating while the group is rebalancing. This // ensures that the coordinator keeps the member in the group for as long // as the duration of the rebalance timeout. If we stop sending heartbeats, // however, then the session timeout may expire before we can rejoin. heartbeat.receiveHeartbeat(time.milliseconds()); } else { heartbeat.failHeartbeat(); // wake up the thread if it's sleeping to reschedule the heartbeat AbstractCoordinator.this.notify(); } } } }); } } } } catch (AuthenticationException e) { log.error("An authentication error occurred in the heartbeat thread", e); this.failed.set(e); } catch (GroupAuthorizationException e) { log.error("A group authorization error occurred in the heartbeat thread", e); this.failed.set(e); } catch (InterruptedException | InterruptException e) { Thread.interrupted(); log.error("Unexpected interrupt received in heartbeat thread", e); this.failed.set(new RuntimeException(e)); } catch (Throwable e) { log.error("Heartbeat thread failed due to unexpected error", e); if (e instanceof RuntimeException) this.failed.set((RuntimeException) e); else this.failed.set(new RuntimeException(e)); } finally { log.debug("Heartbeat thread has closed"); } }
在心跳執行緒中這裡麵包含兩個最重要的超時函式,它們是sessionTimeoutExpired和pollTimeoutExpired。
public boolean sessionTimeoutExpired(long now) { return now - Math.max(lastSessionReset, lastHeartbeatReceive) > sessionTimeoutMs; } public boolean pollTimeoutExpired(long now) { return now - lastPoll > maxPollIntervalMs;
}
2.2.1 sessionTimeoutExpired
如果是sessionTimeout超時,則會被標記為當前協調器處理斷開,此時,會將消費者移除,重新分配分割槽和消費者的對應關係。在Kafka Broker Server中,Consumer Group定義了5中(如果算上Unknown,應該是6種狀態)狀態,org.apache.kafka.common.ConsumerGroupState,如下圖所示:
2.2.2 pollTimeoutExpired
如果觸發了poll超時,此時消費者客戶端會退出ConsumerGroup,當再次poll的時候,會重新加入到ConsumerGroup,觸發RebalanceGroup。而KafkaConsumer Client是不會幫我們重複poll的,需要我們自己在實現的消費邏輯中不停的呼叫poll方法。
3.分割槽與消費執行緒
關於消費分割槽與消費執行緒的對應關係,理論上消費執行緒數應該小於等於分割槽數。之前是有這樣一種觀點,一個消費執行緒對應一個分割槽,當消費執行緒等於分割槽數是最大化執行緒的利用率。直接使用KafkaConsumer Client例項,這樣使用確實沒有什麼問題。但是,如果我們有富裕的CPU,其實還可以使用大於分割槽數的執行緒,來提升消費能力,這就需要我們對KafkaConsumer Client例項進行改造,實現消費策略預計算,利用額外的CPU開啟更多的執行緒,來實現消費任務分片。具體實現,留到下一篇部落格,給大家分享《基於Kafka的分散式查詢SQL引擎》。
4.結束語
這篇部落格就和大家分享到這裡,如果大家在研究學習的過程當中有什麼問題,可以加群進行討論或傳送郵件給我,我會盡我所能為您解答,與君共勉!
另外,博主出書了《Kafka並不難學》和《Hadoop大資料探勘從入門到進階實戰》,喜歡的朋友或同學, 可以在公告欄那裡點選購買連結購買博主的書進行學習,在此感謝大家的支援。關注下面公眾號,根據提示,可免費獲取書籍的教學視訊。