Github user blrunner commented on a diff in the pull request:
https://github.com/apache/tajo/pull/482#discussion_r27782992
--- Diff:
tajo-storage/tajo-storage-kafka/src/main/java/org/apache/tajo/storage/kafka/SimpleConsumerManager.java
---
@@ -0,0 +1,314 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.tajo.storage.kafka;
+
+import java.io.IOException;
+import java.net.InetSocketAddress;
+import java.util.ArrayList;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Map;
+import java.util.Random;
+import java.util.Set;
+
+import kafka.api.FetchRequest;
+import kafka.api.FetchRequestBuilder;
+import kafka.api.PartitionOffsetRequestInfo;
+import kafka.cluster.Broker;
+import kafka.common.ErrorMapping;
+import kafka.common.TopicAndPartition;
+import kafka.javaapi.FetchResponse;
+import kafka.javaapi.OffsetResponse;
+import kafka.javaapi.PartitionMetadata;
+import kafka.javaapi.TopicMetadata;
+import kafka.javaapi.TopicMetadataRequest;
+import kafka.javaapi.consumer.SimpleConsumer;
+import kafka.message.MessageAndOffset;
+
+import org.apache.commons.collections.IteratorUtils;
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.tajo.util.NetUtils;
+
+// SimpleConsumerManager is kafka client for KafkaScanner.
+// It's one per partition. Each partition instantiate this class.
+public class SimpleConsumerManager {
+ private static final Log LOG =
LogFactory.getLog(SimpleConsumerManager.class);
+ // TODO: configurable setting.
+ static final int CONSUMER_TIMEOUT = 30000;
+ static final int CONSUMER_BUFFER_SIZE = 64 * 1024;
+ static final int CONSUMER_FETCH_SIZE = 300 * 1024;
+ static final int FETCH_TRY_NUM = 3;
+
+ private SimpleConsumer consumer = null;
+ private List<InetSocketAddress> brokers = new
ArrayList<InetSocketAddress>();
+ private String topic;
+ private int partition;
+ private String clientId;
+ // leader of this partition.
+ private Broker leader;
+
+ public SimpleConsumerManager(String seedBrokers, String topic, int
partition) throws IOException {
+ this.topic = topic;
+ this.partition = partition;
+ // Identifier of simpleConsumer.
+ this.clientId = SimpleConsumerManager.getIdentifier();
+ this.brokers = SimpleConsumerManager.getBrokerList(seedBrokers);
+ this.leader = findLeader(topic, partition);
+ // consumer creation fail.
+ if (null == leader) {
+ throw new IOException("consumer creation fail");
+ } else {
+ consumer = new SimpleConsumer(leader.host(), leader.port(),
CONSUMER_TIMEOUT, CONSUMER_BUFFER_SIZE, clientId);
+ }
+ }
+
+ /**
+ * Create SimpleConsumer instance. seedBrokers is connection info of
kafka
+ * brokers. ex) localhost:9092,localhost:9091 topic is topic name.
partition
+ * is partition id.
+ *
+ * @param seedBrokers
+ * @param topic
+ * @param partition
+ * @return
+ * @throws IOException
+ */
+ static public SimpleConsumerManager getSimpleConsumerManager(String
seedBrokers, String topic, int partition)
+ throws IOException {
+ return new SimpleConsumerManager(seedBrokers, topic, partition);
+ }
+
+ /**
+ * Return partition ID list of specific topic. Check for seedBrokers.
+ * seedBrokers is kafka brokers.
+ *
+ * @param seedBrokers
+ * @param topic
+ * @return
+ * @throws IOException
+ */
+ static public Set<Integer> getPartitions(String seedBrokers, String
topic) throws IOException {
+ Set<Integer> partitions = new HashSet<Integer>();
+ for (InetSocketAddress seed :
SimpleConsumerManager.getBrokerList(seedBrokers)) {
+ SimpleConsumer consumer = null;
+ try {
+ consumer = new SimpleConsumer(seed.getHostName(), seed.getPort(),
CONSUMER_TIMEOUT, CONSUMER_BUFFER_SIZE,
+ SimpleConsumerManager.getIdentifier() + "partitionLookup");
+ List<String> topics = new ArrayList<String>();
+ topics.add(topic);
+ TopicMetadataRequest req = new TopicMetadataRequest(topics);
+ kafka.javaapi.TopicMetadataResponse resp = consumer.send(req);
+ // call to topicsMetadata() asks the Broker you are connected to
for all
+ // the details about the topic we are interested in
+ List<TopicMetadata> metaData = resp.topicsMetadata();
+ // loop on partitionsMetadata iterates through all the partitions
until
+ // we find the one we want.
+ for (TopicMetadata item : metaData) {
+ for (PartitionMetadata part : item.partitionsMetadata()) {
+ partitions.add(part.partitionId());
+ }
+ }
+ } catch (Exception e) {
+ LOG.error(e.getMessage(), e);
+ throw new IOException(e);
+ } finally {
+ if (consumer != null)
+ consumer.close();
+ }
+ }
+ return partitions;
+ }
+
+ static private List<InetSocketAddress> getBrokerList(String brokers) {
+ List<InetSocketAddress> brokerList = new
ArrayList<InetSocketAddress>();
+ for (String broker : brokers.split(",")) {
+ brokerList.add(NetUtils.createUnresolved(broker));
+ }
+ return brokerList;
+ }
+
+ /**
+ * Create identifier for SimpleConsumer. The SimpleConsumer connects at
kafka
+ * using this identifier.
+ *
+ * @return
+ */
+ static private String getIdentifier() {
+ Random r = new Random();
+ return r.nextLong() + "_" + System.currentTimeMillis();
+ }
+
+ synchronized public void close() {
+ if (null != consumer) {
+ consumer.close();
+ }
+ consumer = null;
+ }
+
+ /**
+ * Fetch data from kafka, as much as 'CONSUMER_FETCH_SIZE' size from
offset.
+ *
+ * @param offset
+ * @return
+ */
+ @SuppressWarnings("unchecked")
+ public List<MessageAndOffset> fetch(long offset) {
+ List<MessageAndOffset> returnData = null;
+ FetchRequest req = new FetchRequestBuilder().clientId(clientId)
+ .addFetch(topic, partition, offset, CONSUMER_FETCH_SIZE).build();
+ if (null != consumer) {
+ FetchResponse fetchResponse = null;
+ // If that fails, find new leader of partition and try again.
+ for (int i = 0; i < FETCH_TRY_NUM; i++) {
+ fetchResponse = consumer.fetch(req);
+ if (fetchResponse.hasError()) {
+ short code = fetchResponse.errorCode(topic, partition);
+ LOG.error("Error fetching data from the Broker:" + leader + "
Reason: " + code + " Try: " + i);
+ if (ErrorMapping.LeaderNotAvailableCode() == code ||
ErrorMapping.NotLeaderForPartitionCode() == code
+ || ErrorMapping.RequestTimedOutCode() == code ||
ErrorMapping.BrokerNotAvailableCode() == code
+ || ErrorMapping.ReplicaNotAvailableCode() == code) {
+ Broker newLeader = findNewLeader();
+ if (null != newLeader) {
+ synchronized (consumer) {
+ this.leader = newLeader;
+ consumer.close();
+ consumer = new SimpleConsumer(leader.host(),
leader.port(), CONSUMER_TIMEOUT, CONSUMER_BUFFER_SIZE,
+ clientId);
+ }
+ }
+ }
+ } else {
+ break;
+ }
+ fetchResponse = null;
+ }
+ if (null != fetchResponse) {
+ Iterator<MessageAndOffset> messages =
fetchResponse.messageSet(topic, partition).iterator();
+ returnData = IteratorUtils.toList(messages);
+ }
+ }
+ return returnData;
+ }
+
+ public long getReadOffset(long whichTime) throws IOException {
+ TopicAndPartition topicAndPartition = new TopicAndPartition(topic,
partition);
+ Map<TopicAndPartition, PartitionOffsetRequestInfo> requestInfo = new
HashMap<TopicAndPartition, PartitionOffsetRequestInfo>();
+ requestInfo.put(topicAndPartition, new
PartitionOffsetRequestInfo(whichTime, 1));
+ kafka.javaapi.OffsetRequest request = new
kafka.javaapi.OffsetRequest(requestInfo,
+ kafka.api.OffsetRequest.CurrentVersion(), clientId);
+ OffsetResponse response = consumer.getOffsetsBefore(request);
+
+ if (response.hasError()) {
+ LOG.error("Error fetching data Offset Data the Broker. Reason: " +
response.errorCode(topic, partition));
+ throw new IOException("Error fetching data Offset Data the Broker");
+ }
+ long[] offsets = response.offsets(topic, partition);
+ return offsets[0];
+ }
+
+ /**
+ * find new leader of partition, if old leader fail.
+ *
+ * @return
+ */
+ synchronized private Broker findNewLeader() {
+ // retry for 3 time.
+ for (int i = 0; i < 3; i++) {
+ boolean goToSleep = false;
+ Broker newLeader = findLeader(topic, partition);
+ if (leader == null) {
+ goToSleep = true;
+ } else if (leader.host().equalsIgnoreCase(newLeader.host()) &&
leader.port() == newLeader.port() && i == 0) {
+ // first time through if the leader hasn't changed give ZooKeeper a
+ // second to recover
+ // second time, assume the broker did recover before failover, or
it was
+ // a non-Broker issue
+ goToSleep = true;
+ } else {
+ return newLeader;
+ }
+ if (goToSleep) {
+ try {
+ Thread.sleep(1000);
+ } catch (InterruptedException ie) {
+ }
+ }
+ }
+ // Unable to find new leader after Broker failure.
+ return null;
+ }
+
+ /**
+ * Find leader broker of specific topic and partition
+ *
+ * @param topic
+ * @param partition
+ * @return
+ */
+ synchronized private Broker findLeader(String topic, int partition) {
--- End diff --
Could you explain why you use synchronized keyword?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
