djkevincr commented on a change in pull request #179: GORA-532: Apache Gora Benchmark initial pull request for review and comments URL: https://github.com/apache/gora/pull/179#discussion_r315624050
########## File path: gora-benchmark/src/main/java/org/apache/gora/benchmark/GoraBenchmarkClient.java ########## @@ -0,0 +1,271 @@ +/** + * 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.gora.benchmark; + +import java.util.Arrays; +import java.util.HashMap; +import java.util.List; +import java.util.Map; +import java.util.Properties; +import java.util.Set; +import java.util.Vector; +import org.apache.gora.query.Query; +import org.apache.gora.query.Result; +import org.apache.gora.store.DataStore; +import org.apache.gora.store.DataStoreFactory; +import org.apache.gora.util.GoraException; +import org.apache.hadoop.conf.Configuration; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; +import com.yahoo.ycsb.ByteIterator; +import com.yahoo.ycsb.DB; +import com.yahoo.ycsb.DBException; +import com.yahoo.ycsb.Status; +import com.yahoo.ycsb.StringByteIterator; +import com.yahoo.ycsb.workloads.CoreWorkload; +import org.apache.gora.benchmark.generated.User; + +/** + * The Class GoraBenchmarkClient + * + * @author sc306 This class extends the Yahoo! Cloud Service Benchmark benchmark + * {@link #com.yahoo.ycsb.DB DB} class to provide functionality for + * {@link #insert(String, String, HashMap) insert}, + * {@link #read(String, String, Set, HashMap) read}, + * {@link #scan(String, String, int, Set, Vector) scan} and + * {@link #update(String, String, HashMap) update} methods as per Apache + * Gora implementation. + */ +public class GoraBenchmarkClient extends DB { + private static final Logger LOG = LoggerFactory.getLogger(GoraBenchmarkClient.class); + private static final String FIELDS[] = User._ALL_FIELDS; + private static volatile boolean executed; + public static int fieldCount; + /** This is only for set to array conversion in {@link read()} method */ + private String[] DUMMY_ARRAY = new String[0]; + DataStore<String, User> dataStore; + GoraBenchmarkUtils goraBenchmarkUtils = new GoraBenchmarkUtils(); + User user = new User(); + private Properties prop; + + public GoraBenchmarkClient() { + } + + /*** + * Initialisation method. This method is called once for each database + * instance. + */ + public void init() throws DBException { + try { + // Get YCSB properties + prop = getProperties(); + fieldCount = Integer + .parseInt(prop.getProperty(CoreWorkload.FIELD_COUNT_PROPERTY, CoreWorkload.FIELD_COUNT_PROPERTY_DEFAULT)); + String keyClass = prop.getProperty("key.class", "java.lang.String"); + String persistentClass = prop.getProperty("persistent.class", "org.apache.gora.benchmark.generated.User"); + Properties p = DataStoreFactory.createProps(); + dataStore = DataStoreFactory.getDataStore(keyClass, persistentClass, p, new Configuration()); + synchronized (GoraBenchmarkClient.class) { Review comment: So basically, when class level locking is used, for multiple instances of the same GoraBenchmarkClient class in single JVM, only one thread will get access to class level monitor ( only a single monitor per class of instances ) and execute synchronized block of a single instance from all the instances available in JVM. In object level synchronization, threads just have to wait for each instance's monitor as opposed to class object monitor. Threads can paralleled acquire instance monitors. Class level synchronization can lead to higher thread contention and possibly higher waiting times for threads. This can be a overhead since if we don't really want to synchronize on all instances of the class and synchronization is only required for per single instance. If you compare two case class level and object level, class level has the highest level synchronization/safety as mentioned but it comes with overhead. Think about two client instances of GoraBenchmarkClient and how these instances are getting accessed/methods being call when a we perform a benchmark. Think about above statements can have a impact. These kind of concurrency related things might influence benchmark results we get. ---------------------------------------------------------------- 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: [email protected] With regards, Apache Git Services
