當前位置: 首頁>>代碼示例>>Java>>正文


Java PartitionInfo.partition方法代碼示例

本文整理匯總了Java中org.apache.kafka.common.PartitionInfo.partition方法的典型用法代碼示例。如果您正苦於以下問題:Java PartitionInfo.partition方法的具體用法?Java PartitionInfo.partition怎麽用?Java PartitionInfo.partition使用的例子?那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在org.apache.kafka.common.PartitionInfo的用法示例。


在下文中一共展示了PartitionInfo.partition方法的11個代碼示例,這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚,您的評價將有助於係統推薦出更棒的Java代碼示例。

示例1: fillInMissingPartitions

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
/**
 * Add empty load of all the partitions that exists in the current cluster but missing from the
 * metric aggregation result.
 */
private void fillInMissingPartitions(Map<TopicPartition, Snapshot[]> loadSnapshots,
                                            Cluster kafkaCluster,
                                            ClusterModel clusterModel) throws ModelInputException {
  // There must be at least one entry, otherwise there will be exception thrown earlier. So we don't need to
  // check if it has next
  Snapshot[] snapshotsForTimestamps = loadSnapshots.values().iterator().next();
  Snapshot[] emptyLoadSnapshots = new Snapshot[snapshotsForTimestamps.length];
  for (int i = 0; i < emptyLoadSnapshots.length; i++) {
    emptyLoadSnapshots[i] = new Snapshot(snapshotsForTimestamps[i].time());
  }
  for (Node node : kafkaCluster.nodes()) {
    for (PartitionInfo partitionInfo : kafkaCluster.partitionsForNode(node.id())) {
      TopicPartition tp = new TopicPartition(partitionInfo.topic(), partitionInfo.partition());
      if (!loadSnapshots.containsKey(tp)) {
        populateSnapshots(kafkaCluster, clusterModel, tp, emptyLoadSnapshots);
      }
    }
  }
}
 
開發者ID:linkedin,項目名稱:cruise-control,代碼行數:24,代碼來源:LoadMonitor.java

示例2: getCount

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
/**
 * Gets the total message count for the topic.
 * <b>WARNING: Don't use with compacted topics</b>
 */
@SuppressWarnings("unchecked")
public long getCount(String kafkaBrokers, String topic) {
    KafkaConsumer consumer = buildConsumer(kafkaBrokers);
    try {
        @SuppressWarnings("unchecked")
        Map<String, List<PartitionInfo>> topics = consumer.listTopics();
        List<PartitionInfo> partitionInfos = topics.get(topic);
        if (partitionInfos == null) {
            logger.warn("Partition information was not found for topic {}", topic);
            return 0;
        } else {
            Collection<TopicPartition> partitions = new ArrayList<>();
            for (PartitionInfo partitionInfo : partitionInfos) {
                TopicPartition partition = new TopicPartition(topic, partitionInfo.partition());
                partitions.add(partition);
            }
            Map<TopicPartition, Long> endingOffsets = consumer.endOffsets(partitions);
            Map<TopicPartition, Long> beginningOffsets = consumer.beginningOffsets(partitions);
            return diffOffsets(beginningOffsets, endingOffsets);
        }
    } finally {
        consumer.close();
    }
}
 
開發者ID:Sixt,項目名稱:ja-micro,代碼行數:29,代碼來源:TopicMessageCounter.java

示例3: hasPartition

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
private boolean hasPartition(final TopicPartition topicPartition) {
    final List<PartitionInfo> partitions = partitionInfo.get(topicPartition.topic());

    if (partitions == null) {
        return false;
    }

    for (final PartitionInfo partition : partitions) {
        if (partition.partition() == topicPartition.partition()) {
            return true;
        }
    }

    return false;

}
 
開發者ID:YMCoding,項目名稱:kafka-0.11.0.0-src-with-comment,代碼行數:17,代碼來源:StoreChangelogReader.java

示例4: getBrokerLeaderPartitions

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
public Map<Integer, List<TopicPartition>> getBrokerLeaderPartitions(
    Map<String, List<PartitionInfo>> topicPartitonInfoMap) {
  Map<Integer, List<TopicPartition>> result = new HashMap<>();

  for (String topic : topicPartitonInfoMap.keySet()) {
    List<PartitionInfo> partitionInfoList = topicPartitonInfoMap.get(topic);
    if (partitionInfoList == null) {
      LOG.error("Failed to get partition info for {}", topic);
      continue;
    }

    for (PartitionInfo info : partitionInfoList) {
      Node leaderNode = info.leader();
      if (leaderNode != null) {
        result.putIfAbsent(leaderNode.id(), new ArrayList<>());
        TopicPartition topicPartiton = new TopicPartition(info.topic(), info.partition());
        result.get(leaderNode.id()).add(topicPartiton);
      }
    }
  }
  return result;
}
 
開發者ID:pinterest,項目名稱:doctorkafka,代碼行數:23,代碼來源:KafkaClusterManager.java

示例5: run

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
/**
 * When an object implementing interface <code>Runnable</code> is used
 * to create a thread, starting the thread causes the object's
 * <code>run</code> method to be called in that separately executing
 * thread.
 * <p>
 * The general contract of the method <code>run</code> is that it may
 * take any action whatsoever.
 *
 * @see Thread#run()
 */
@Override
public void run() {
    String group = "kafka-insight-logOffsetListener";
    int sleepTime = 60000;
    KafkaConsumer<Array<Byte>, Array<Byte>> kafkaConsumer = null;

    while (true) {

        try {
            if (null == kafkaConsumer) {
                kafkaConsumer = KafkaUtils.createNewKafkaConsumer(brokersInfo, group);
            }

            Map<String, List<PartitionInfo>> topicPartitionsMap = kafkaConsumer.listTopics();
            for (List<PartitionInfo> partitionInfoList : topicPartitionsMap.values()) {
                for (PartitionInfo partitionInfo : partitionInfoList) {
                    TopicPartition topicPartition = new TopicPartition(partitionInfo.topic(), partitionInfo.partition());
                    Collection<TopicPartition> topicPartitions = Arrays.asList(topicPartition);
                    kafkaConsumer.assign(topicPartitions);
                    kafkaConsumer.seekToEnd(topicPartitions);
                    Long logEndOffset = kafkaConsumer.position(topicPartition);
                    logEndOffsetMap.put(topicPartition, logEndOffset);
                }
            }

            Thread.sleep(sleepTime);

        } catch (Exception e) {
            e.printStackTrace();
            if (null != kafkaConsumer) {
                kafkaConsumer.close();
                kafkaConsumer = null;
            }
        }
    }

}
 
開發者ID:dubin555,項目名稱:Kafka-Insight,代碼行數:49,代碼來源:KafkaOffsetGetter.java

示例6: getKafkaOffsets

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
private Map<TopicPartition, OffsetAndMetadata> getKafkaOffsets(
    KafkaConsumer<String, byte[]> client, String topicStr) {
  Map<TopicPartition, OffsetAndMetadata> offsets = new HashMap<>();
  List<PartitionInfo> partitions = client.partitionsFor(topicStr);
  for (PartitionInfo partition : partitions) {
    TopicPartition key = new TopicPartition(topicStr, partition.partition());
    OffsetAndMetadata offsetAndMetadata = client.committed(key);
    if (offsetAndMetadata != null) {
      offsets.put(key, offsetAndMetadata);
    }
  }
  return offsets;
}
 
開發者ID:moueimei,項目名稱:flume-release-1.7.0,代碼行數:14,代碼來源:KafkaSource.java

示例7: getKafkaOffsets

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
private Map<TopicPartition, OffsetAndMetadata> getKafkaOffsets(
    KafkaConsumer<String, byte[]> client) {
  Map<TopicPartition, OffsetAndMetadata> offsets = new HashMap<>();
  List<PartitionInfo> partitions = client.partitionsFor(topicStr);
  for (PartitionInfo partition : partitions) {
    TopicPartition key = new TopicPartition(topicStr, partition.partition());
    OffsetAndMetadata offsetAndMetadata = client.committed(key);
    if (offsetAndMetadata != null) {
      offsets.put(key, offsetAndMetadata);
    }
  }
  return offsets;
}
 
開發者ID:moueimei,項目名稱:flume-release-1.7.0,代碼行數:14,代碼來源:KafkaChannel.java

示例8: resetCorrectOffsets

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
/**
 * 按上次記錄重置offsets
 */
private void resetCorrectOffsets() {
    consumer.pause(consumer.assignment());
    Map<String, List<PartitionInfo>> topicInfos = consumer.listTopics();
    Set<String> topics = topicInfos.keySet();

    List<String> expectTopics = new ArrayList<>(topicHandlers.keySet());

    List<PartitionInfo> patitions = null;
    for (String topic : topics) {
        if (!expectTopics.contains(topic))
            continue;

        patitions = topicInfos.get(topic);
        for (PartitionInfo partition : patitions) {
            try {
                //期望的偏移
                long expectOffsets = consumerContext.getLatestProcessedOffsets(topic,
                    partition.partition());
                //
                TopicPartition topicPartition = new TopicPartition(topic,
                    partition.partition());
                OffsetAndMetadata metadata = consumer
                    .committed(new TopicPartition(partition.topic(), partition.partition()));
                if (expectOffsets >= 0) {
                    if (expectOffsets < metadata.offset()) {
                        consumer.seek(topicPartition, expectOffsets);
                        logger.info("seek Topic[{}] partition[{}] from {} to {}", topic,
                            partition.partition(), metadata.offset(), expectOffsets);
                    }
                }
            } catch (Exception e) {
                logger.warn("try seek topic[" + topic + "] partition[" + partition.partition()
                            + "] offsets error");
            }
        }
    }
    consumer.resume(consumer.assignment());
}
 
開發者ID:warlock-china,項目名稱:azeroth,代碼行數:42,代碼來源:NewApiTopicConsumer.java

示例9: recentSnapshotsForTopic

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
/**
 * Get all the snapshots for a topic before the given time, excluding the snapshot window that the given time falls in.
 */
private MetricSampleAggregationResult recentSnapshotsForTopic(String topic,
                                                              Cluster cluster,
                                                              Map<Long, Map<TopicPartition, AggregatedMetrics>> readyWindowedAggMetrics,
                                                              boolean includeAllTopics)
    throws NotEnoughSnapshotsException {
  List<PartitionInfo> partitions = cluster.partitionsForTopic(topic);
  MetricSampleAggregationResult result = new MetricSampleAggregationResult(currentGeneration(), includeAllTopics);

  for (PartitionInfo partition : partitions) {
    TopicPartition tp = new TopicPartition(topic, partition.partition());
    Snapshot[] snapshots = new Snapshot[readyWindowedAggMetrics.size()];
    int index = 0;
    for (Map.Entry<Long, Map<TopicPartition, AggregatedMetrics>> entry : readyWindowedAggMetrics.entrySet()) {
      long snapshotWindow = entry.getKey();
      SnapshotAndImputation partitionSnapShot =
              partitionSnapshotForWindow(tp, cluster, snapshotWindow, snapshotWindow, includeAllTopics, true);

      // There is an invalid partition for the topic, we just exclude the topic by returning the result immediately.
      // At this point the result does not contain any snapshot, but only partition flaw.
      if (partitionSnapShot.snapshot() == null) {
        result.discardAllSnapshots();
        result.recordPartitionWithSampleFlaw(tp, snapshotWindow, partitionSnapShot.imputation());
        return result;
      }

      // Record the imputation and return the result without snapshot if there has been too many imputations.
      if (partitionSnapShot.imputation() != null) {
        result.recordPartitionWithSampleFlaw(tp, snapshotWindow, partitionSnapShot.imputation());
        if (result.sampleFlaw(tp).size() > _numSnapshots * MAX_SNAPSHOT_PERCENT_WITH_FLAWS) {
          LOG.debug("{} already has {} snapshot with flaws, excluding the partition from the model.",
                    tp, result.sampleFlaw(tp).size());
          result.discardAllSnapshots();
          return result;
        }
      }

      snapshots[index++] = partitionSnapShot.snapshot();
    }
    result.addPartitionSnapshots(new TopicPartition(partition.topic(), partition.partition()), snapshots);
  }
  return result;
}
 
開發者ID:linkedin,項目名稱:cruise-control,代碼行數:46,代碼來源:MetricSampleAggregator.java

示例10: getUnderReplicatedPartitions

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
/**
 * Call the kafka api to get the list of under-replicated partitions.
 * When a topic partition loses all of its replicas, it will not have a leader broker.
 * We need to handle this special case in detecting under replicated topic partitions.
 */
public static List<PartitionInfo> getUnderReplicatedPartitions(
    String zkUrl, List<String> topics,
    scala.collection.mutable.Map<String, scala.collection.Map<Object, Seq<Object>>>
        partitionAssignments,
    Map<String, Integer> replicationFactors,
    Map<String, Integer> partitionCounts) {
  List<PartitionInfo> underReplicated = new ArrayList();
  KafkaConsumer kafkaConsumer = KafkaUtils.getKafkaConsumer(zkUrl);
  for (String topic : topics) {
    List<PartitionInfo> partitionInfoList = kafkaConsumer.partitionsFor(topic);
    if (partitionInfoList == null) {
      LOG.error("Failed to get partition info for {}", topic);
      continue;
    }
    int numPartitions = partitionCounts.get(topic);

    // when a partition loses all replicas and does not have a live leader,
    // kafkaconsumer.partitionsFor(...) will not return info for that partition.
    // the noLeaderFlag array is used to detect partitions that have no leaders
    boolean[] noLeaderFlags = new boolean[numPartitions];
    for (int i = 0; i < numPartitions; i++) {
      noLeaderFlags[i] = true;
    }
    for (PartitionInfo info : partitionInfoList) {
      if (info.inSyncReplicas().length < info.replicas().length &&
          replicationFactors.get(info.topic()) > info.inSyncReplicas().length) {
        underReplicated.add(info);
      }
      noLeaderFlags[info.partition()] = false;
    }

    // deal with the partitions that do not have leaders
    for (int partitionId = 0; partitionId < numPartitions; partitionId++) {
      if (noLeaderFlags[partitionId]) {
        Seq<Object> seq = partitionAssignments.get(topic).get().get(partitionId).get();
        Node[] nodes = JavaConverters.seqAsJavaList(seq).stream()
            .map(val -> new Node((Integer) val, "", -1)).toArray(Node[]::new);
        PartitionInfo partitionInfo =
            new PartitionInfo(topic, partitionId, null, nodes, new Node[0]);
        underReplicated.add(partitionInfo);
      }
    }
  }
  return underReplicated;
}
 
開發者ID:pinterest,項目名稱:doctorkafka,代碼行數:51,代碼來源:KafkaClusterManager.java

示例11: OutOfSyncReplica

import org.apache.kafka.common.PartitionInfo; //導入方法依賴的package包/類
public OutOfSyncReplica(PartitionInfo partitionInfo) {
  this.topicPartition = new TopicPartition(partitionInfo.topic(), partitionInfo.partition());
  this.inSyncBrokers = getInSyncReplicas(partitionInfo);
  this.outOfSyncBrokers = getOutOfSyncReplicas(partitionInfo);
  this.leader = partitionInfo.leader();
}
 
開發者ID:pinterest,項目名稱:doctorkafka,代碼行數:7,代碼來源:OutOfSyncReplica.java


注:本文中的org.apache.kafka.common.PartitionInfo.partition方法示例由純淨天空整理自Github/MSDocs等開源代碼及文檔管理平台,相關代碼片段篩選自各路編程大神貢獻的開源項目,源碼版權歸原作者所有,傳播和使用請參考對應項目的License;未經允許,請勿轉載。