Clarkkkkk commented on a change in pull request #7550: [FLINK-10245] [Streaming Connector] Add Pojo, Tuple, Row and Scala Product DataStream Sink and Upsert Table Sink for HBase URL: https://github.com/apache/flink/pull/7550#discussion_r299848380
########## File path: flink-connectors/flink-hbase/src/main/java/org/apache/flink/streaming/connectors/hbase/HBaseSinkFunctionBase.java ########## @@ -0,0 +1,332 @@ +/* + * 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.flink.streaming.connectors.hbase; + +import org.apache.flink.api.common.typeinfo.TypeInformation; +import org.apache.flink.configuration.Configuration; +import org.apache.flink.runtime.concurrent.FutureUtils; +import org.apache.flink.runtime.state.FunctionInitializationContext; +import org.apache.flink.runtime.state.FunctionSnapshotContext; +import org.apache.flink.streaming.api.checkpoint.CheckpointedFunction; +import org.apache.flink.streaming.api.functions.sink.RichSinkFunction; +import org.apache.flink.streaming.connectors.hbase.util.HBaseUtils; + +import org.apache.hadoop.hbase.client.Append; +import org.apache.hadoop.hbase.client.Connection; +import org.apache.hadoop.hbase.client.Delete; +import org.apache.hadoop.hbase.client.HTable; +import org.apache.hadoop.hbase.client.Increment; +import org.apache.hadoop.hbase.client.Mutation; +import org.apache.hadoop.hbase.client.Put; +import org.apache.hadoop.hbase.client.Table; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import java.util.LinkedList; +import java.util.List; +import java.util.Map; +import java.util.concurrent.CompletableFuture; +import java.util.concurrent.ExecutionException; +import java.util.concurrent.ExecutorService; +import java.util.concurrent.Executors; +import java.util.concurrent.ScheduledThreadPoolExecutor; +import java.util.concurrent.TimeUnit; +import java.util.concurrent.TimeoutException; + +/** + * HBaseSinkFunctionBase is the common abstract class of {@link HBasePojoSinkFunction}, {@link HBaseTupleSinkFunction}, + * {@link HBaseScalaProductSinkFunction} and {@link HBaseUpsertTableSink.HBaseUpsertSinkFunction}. + * + * @param <IN> Type of the elements emitted by this sink + */ +public abstract class HBaseSinkFunctionBase<IN> extends RichSinkFunction<IN> implements CheckpointedFunction { + + private static final Logger log = LoggerFactory.getLogger(HBaseSinkFunctionBase.class); + + // ------------------------------------------------------------------------ + // Internal bulk processor configuration + // ------------------------------------------------------------------------ + + public static final String CONFIG_KEY_BATCH_FLUSH_ENABLE = "connector.batch-flush.enable"; + public static final String CONFIG_KEY_BATCH_FLUSH_MAX_MUTATIONS = "connector.batch-flush.max-mutations"; + public static final String CONFIG_KEY_BATCH_FLUSH_MAX_SIZE_MB = "connector.batch-flush.max-size.mb"; + public static final String CONFIG_KEY_BATCH_FLUSH_INTERVAL_MS = "connector.batch-flush.interval.ms"; + public static final String CONFIG_KEY_BATCH_FLUSH_MAX_RETRIES = "connector.batch-flush.backoff.max-retries"; + public static final String CONFIG_KEY_BATCH_FLUSH_MAX_TIMEOUT_MS = "connector.batch-flush.max-timeout"; + + private final Map<String, String> userConfig; + + protected final int rowKeyIndex; + protected final String[] fieldNames; + protected final TypeInformation<?>[] fieldTypes; + protected final String[] columnFamilies; + protected final String[] qualifiers; + protected final int[] fieldElementIndexMapping; + + /** The timer that triggers periodic flush to HBase. */ + private ScheduledThreadPoolExecutor executor; + + private ExecutorService flushExecutor; + + /** The lock to safeguard the flush commits. */ + private transient Object lock; + + private Connection connection; + private transient Table hTable; + private boolean isBuildingTable = false; + + private HBaseClientWrapper client; + private List<Mutation> mutaionBuffer = new LinkedList<>(); + private long estimateSize = 0; + + private final boolean batchFlushEnable; + private long batchFlushMaxMutations; + private long batchFlushMaxSizeInBits; + private long batchFlushIntervalMillis; + private int batchFlushMaxRetries; + private long batchFlushMaxTimeoutMillis; + + private boolean isRunning = false; + + public HBaseSinkFunctionBase( + String clusterKey, + String tableName, + Map<String, String> userConfig, + int rowKeyIndex, + String[] outputFieldNames, + String[] fieldNames, + String[] columnFamilies, + String[] qualifiers, + TypeInformation<?>[] fieldTypes) { + this(null, userConfig, rowKeyIndex, outputFieldNames, fieldNames, columnFamilies, qualifiers, fieldTypes); + this.client = new HBaseClientWrapper().clusterKey(clusterKey).tableName(tableName); + this.isBuildingTable = true; + } + + public HBaseSinkFunctionBase( + Table hTable, + Map<String, String> userConfig, + int rowKeyIndex, + String[] outputFieldNames, + String[] fieldNames, + String[] columnFamilies, + String[] qualifiers, + TypeInformation<?>[] fieldTypes){ + this.connection = null; + this.hTable = hTable; + this.userConfig = userConfig; + this.rowKeyIndex = rowKeyIndex; + this.fieldNames = fieldNames; + this.columnFamilies = columnFamilies; + this.qualifiers = qualifiers; + this.fieldTypes = fieldTypes; + + int[] fieldElementIndexMapping = new int[fieldNames.length]; + for (int i = 0; i < fieldNames.length; i++) { + fieldElementIndexMapping[i] = -1; + for (int j = 0; j < outputFieldNames.length; j++) { + if (fieldNames[i].equals(outputFieldNames[j])) { + fieldElementIndexMapping[i] = j; + break; + } + } + if (fieldElementIndexMapping[i] == -1) { + throw new RuntimeException("The field " + outputFieldNames[i] + " is not found in the result stream."); + } + } + this.fieldElementIndexMapping = fieldElementIndexMapping; + + batchFlushEnable = userConfig.getOrDefault(CONFIG_KEY_BATCH_FLUSH_ENABLE, "false").equals("true"); + batchFlushMaxMutations = Long.parseLong(userConfig.getOrDefault(CONFIG_KEY_BATCH_FLUSH_MAX_MUTATIONS, "128")); + batchFlushMaxSizeInBits = Long.parseLong(userConfig.getOrDefault(CONFIG_KEY_BATCH_FLUSH_MAX_SIZE_MB, "2")) * 1024 * 1024 * 8; + batchFlushIntervalMillis = Long.parseLong(userConfig.getOrDefault(CONFIG_KEY_BATCH_FLUSH_INTERVAL_MS, "1000")); + batchFlushMaxRetries = Integer.parseInt(userConfig.getOrDefault(CONFIG_KEY_BATCH_FLUSH_MAX_RETRIES, "3")); + batchFlushMaxTimeoutMillis = Long.parseLong(userConfig.getOrDefault(CONFIG_KEY_BATCH_FLUSH_MAX_TIMEOUT_MS, "5000")); + } + + @Override + public void open(Configuration configuration) throws Exception { + this.lock = new Object(); + if (isBuildingTable) { + if (client != null) { + this.connection = client.buildConnection(); + } + if (connection != null) { + this.hTable = client.buildTable(connection); + } + } + if (hTable == null) { + throw new RuntimeException("Cannot build connection for hbase sink, please check the configuraiton."); + } + if (batchFlushEnable) { + ((HTable) hTable).setAutoFlush(false, false); + } else { + ((HTable) hTable).setAutoFlush(true, false); + } + this.executor = new ScheduledThreadPoolExecutor(1); + this.flushExecutor = Executors.newFixedThreadPool(3); + if (batchFlushEnable && batchFlushIntervalMillis > 0) { + executor.scheduleAtFixedRate(() -> { + if (this.hTable != null && this.hTable instanceof HTable) { + synchronized (lock) { + try { + flushToHBaseWithRetryAndTimeout(); + } catch (Exception e){ + log.warn("Scheduled flush operation to HBase cannot be finished.", e); + } + } + } + }, batchFlushIntervalMillis, batchFlushIntervalMillis, TimeUnit.MILLISECONDS); + } + isRunning = true; + } + + @Override + public void invoke(IN value, Context context) throws Exception { + Mutation mutation = extract(value); + long mutationSize = mutation.heapSize(); + if (batchFlushEnable) { + if (estimateSize != 0 && (estimateSize + mutationSize > batchFlushMaxSizeInBits || mutaionBuffer.size() + 1 > batchFlushMaxMutations)) { + synchronized (lock){ + if (estimateSize != 0 && (estimateSize + mutationSize > batchFlushMaxSizeInBits Review comment: After discussion with Hequn Cheng in Jira ticket, I think we might be able to reduce some writes by pre computation with the mutator list. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
