src/main/java/org/apache/giraph/graph/partition/HashMasterPartitioner.java

/*
 * 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.giraph.graph.partition;

import java.util.ArrayList;
import java.util.Collection;
import java.util.Iterator;
import java.util.List;

import org.apache.giraph.graph.WorkerInfo;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.io.Writable;
import org.apache.hadoop.io.WritableComparable;
import org.apache.log4j.Logger;

/**
 * Master will execute a hash based partitioning.
 *
 * @param <I> Vertex index value
 * @param <V> Vertex value
 * @param <E> Edge value
 * @param <M> Message value
 */
@SuppressWarnings("rawtypes")
public class HashMasterPartitioner<I extends WritableComparable,
        V extends Writable, E extends Writable, M extends Writable> implements
        MasterGraphPartitioner<I, V, E, M> {
    /** Provided configuration */
    private Configuration conf;
    /** Specified partition count (overrides calculation) */
    private final int userPartitionCount;
    /** Partition count (calculated in createInitialPartitionOwners) */
    private int partitionCount = -1;
    /** Save the last generated partition owner list */
    private List<PartitionOwner> partitionOwnerList;
    /** Class logger */
    private static Logger LOG = Logger.getLogger(HashMasterPartitioner.class);

    /**
     * ZooKeeper has a limit of the data in a single znode of 1 MB and
     * each entry can go be on the average somewhat more than 300 bytes
     */
    private static final int MAX_PARTTIONS = 1024 * 1024 / 350;

    /**
     * Multiplier for the current workers squared
     */
    public static final String PARTITION_COUNT_MULTIPLIER =
        "hash.masterPartitionCountMultipler";
    public static final float DEFAULT_PARTITION_COUNT_MULTIPLIER = 1.0f;

    /** Overrides default partition count calculation if not -1 */
    public static final String USER_PARTITION_COUNT =
        "hash.userPartitionCount";
    public static final int DEFAULT_USER_PARTITION_COUNT = -1;

    public HashMasterPartitioner(Configuration conf) {
        this.conf = conf;
        userPartitionCount = conf.getInt(USER_PARTITION_COUNT,
                                         DEFAULT_USER_PARTITION_COUNT);
    }

    @Override
    public Collection<PartitionOwner> createInitialPartitionOwners(
            Collection<WorkerInfo> availableWorkerInfos, int maxWorkers) {
        if (availableWorkerInfos.isEmpty()) {
            throw new IllegalArgumentException(
                "createInitialPartitionOwners: No available workers");
        }
        List<PartitionOwner> ownerList = new ArrayList<PartitionOwner>();
        Iterator<WorkerInfo> workerIt = availableWorkerInfos.iterator();
        if (userPartitionCount == DEFAULT_USER_PARTITION_COUNT) {
            float multiplier = conf.getFloat(
                PARTITION_COUNT_MULTIPLIER,
                DEFAULT_PARTITION_COUNT_MULTIPLIER);
            partitionCount =
                Math.max((int) (multiplier * availableWorkerInfos.size() *
                         availableWorkerInfos.size()),
                         1);
        } else {
            partitionCount = userPartitionCount;
        }
        if (LOG.isInfoEnabled()) {
            LOG.info("createInitialPartitionOwners: Creating " +
                     partitionCount + ", default would have been " +
                     (availableWorkerInfos.size() *
                      availableWorkerInfos.size()) + " partitions.");
        }
        if (partitionCount > MAX_PARTTIONS) {
            LOG.warn("createInitialPartitionOwners: " +
                    "Reducing the partitionCount to " + MAX_PARTTIONS +
                    " from " + partitionCount);
            partitionCount = MAX_PARTTIONS;
        }

        for (int i = 0; i < partitionCount; ++i) {
            PartitionOwner owner = new BasicPartitionOwner(i, workerIt.next());
            if (!workerIt.hasNext()) {
                workerIt = availableWorkerInfos.iterator();
            }
            ownerList.add(owner);
        }
        this.partitionOwnerList = ownerList;
        return ownerList;
    }


    @Override
    public Collection<PartitionOwner> getCurrentPartitionOwners() {
        return partitionOwnerList;
    }

    /**
     * Subclasses can set the partition owner list.
     *
     * @param partitionOwnerList New partition owner list.
     */
    protected void setPartitionOwnerList(List<PartitionOwner>
            partitionOwnerList) {
        this.partitionOwnerList = partitionOwnerList;
    }

    @Override
    public Collection<PartitionOwner> generateChangedPartitionOwners(
            Collection<PartitionStats> allPartitionStatsList,
            Collection<WorkerInfo> availableWorkerInfos,
            int maxWorkers,
            long superstep) {
        return PartitionBalancer.balancePartitionsAcrossWorkers(
            conf,
            partitionOwnerList,
            allPartitionStatsList,
            availableWorkerInfos);
    }

    @Override
    public PartitionStats createPartitionStats() {
        return new PartitionStats();
    }

}