org.apache.drill.exec.planner.fragment.HardAffinityFragmentParallelizer Maven / Gradle / Ivy

Go to download
/*
 * 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.drill.exec.planner.fragment;

import org.apache.drill.shaded.guava.com.google.common.collect.Lists;
import org.apache.drill.shaded.guava.com.google.common.collect.Maps;
import org.apache.drill.exec.physical.EndpointAffinity;
import org.apache.drill.exec.physical.PhysicalOperatorSetupException;
import org.apache.drill.exec.proto.CoordinationProtos.DrillbitEndpoint;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.util.Collection;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;

/**
 * Implementation of {@link FragmentParallelizer} where fragment requires
 * running on a given set of endpoints. Width per node is depended on the
 * affinity to the endpoint and total width (calculated using costs).
 */
public class HardAffinityFragmentParallelizer implements FragmentParallelizer {
  private static final Logger logger = LoggerFactory.getLogger(HardAffinityFragmentParallelizer.class);

  public static final HardAffinityFragmentParallelizer INSTANCE = new HardAffinityFragmentParallelizer();

  private static String EOL = System.getProperty("line.separator");

  private HardAffinityFragmentParallelizer() { /* singleton */}

  @Override
  public void parallelizeFragment(final Wrapper fragmentWrapper, final ParallelizationParameters parameters,
      final Collection activeEndpoints) throws PhysicalOperatorSetupException {

    final Stats stats = fragmentWrapper.getStats();
    final ParallelizationInfo pInfo = stats.getParallelizationInfo();

    int totalMaxWidth = 0;

    // Go through the affinity map and extract the endpoints that have mandatory assignment requirement
    final Map endpointPool = Maps.newHashMap();
    for (Entry entry : pInfo.getEndpointAffinityMap().entrySet()) {
      if (entry.getValue().isAssignmentRequired()) {
        endpointPool.put(entry.getKey(), entry.getValue());

        // Limit the max width of the endpoint to allowed max width.
        totalMaxWidth += Math.min(parameters.getMaxWidthPerNode(), entry.getValue().getMaxWidth());
        if (totalMaxWidth < 0) {
          // If the totalWidth overflows, just keep it at the max value.
          totalMaxWidth = Integer.MAX_VALUE;
        }
      }
    }

    // Step 1: Find the width taking into account various parameters
    // 1.1. Find the parallelization based on cost. Use max cost of all operators in this fragment; this is consistent
    //      with the calculation that ExcessiveExchangeRemover uses.
    int width = (int) Math.ceil(stats.getMaxCost() / parameters.getSliceTarget());

    // 1.2. Make sure the width is at least the number of endpoints that require an assignment
    width = Math.max(endpointPool.size(), width);

    // 1.3. Cap the parallelization width by fragment level width limit and system level per query width limit
    width = Math.max(1, Math.min(width, pInfo.getMaxWidth()));
    checkOrThrow(endpointPool.size() <= width, logger,
        "Number of mandatory endpoints ({}) that require an assignment is more than the allowed fragment max " +
            "width ({}).", endpointPool.size(), pInfo.getMaxWidth());

    // 1.4 Cap the parallelization width by global max query width
    width = Math.max(1, Math.min(width, parameters.getMaxGlobalWidth()));
    checkOrThrow(endpointPool.size() <= width, logger,
        "Number of mandatory endpoints ({}) that require an assignment is more than the allowed global query " +
            "width ({}).", endpointPool.size(), parameters.getMaxGlobalWidth());

    // 1.5 Cap the parallelization width by max allowed parallelization per node
    width = Math.max(1, Math.min(width, endpointPool.size() * parameters.getMaxWidthPerNode()));

    // 1.6 Cap the parallelization width by total of max allowed width per node. The reason is if the width is more,
    // we end up allocating more work units to one or more endpoints that don't have that many work units.
    width = Math.min(totalMaxWidth, width);

    // Step 2: Select the endpoints
    final Map endpoints = Maps.newHashMap();

    // 2.1 First add each endpoint from the pool once so that the mandatory assignment requirement is fulfilled.
    for (Entry entry : endpointPool.entrySet()) {
      endpoints.put(entry.getKey(), 1);
    }
    int totalAssigned = endpoints.size();

    // 2.2 Assign the remaining slots to endpoints proportional to the affinity of each endpoint
    int remainingSlots = width - endpoints.size();
    while (remainingSlots > 0) {
      for (EndpointAffinity epAf : endpointPool.values()) {
        final int moreAllocation = (int) Math.ceil(epAf.getAffinity() * remainingSlots);
        int currentAssignments = endpoints.get(epAf.getEndpoint());
        for (int i=0;
             i < moreAllocation &&
                totalAssigned < width &&
                currentAssignments < parameters.getMaxWidthPerNode() &&
                currentAssignments < epAf.getMaxWidth();
             i++) {
          totalAssigned++;
          currentAssignments++;
        }
        endpoints.put(epAf.getEndpoint(), currentAssignments);
      }
      final int previousRemainingSlots = remainingSlots;
      remainingSlots = width - totalAssigned;
      if (previousRemainingSlots == remainingSlots) {
        logger.error("Can't parallelize fragment: " +
            "Every mandatory node has exhausted the maximum width per node limit." + EOL +
            "Endpoint pool: {}" + EOL + "Assignment so far: {}" + EOL + "Width: {}", endpointPool, endpoints, width);
        throw new PhysicalOperatorSetupException("Can not parallelize fragment.");
      }
    }

    final List assignedEndpoints = Lists.newArrayList();
    for(Entry entry : endpoints.entrySet()) {
      for(int i=0; i < entry.getValue(); i++) {
        assignedEndpoints.add(entry.getKey());
      }
    }

    fragmentWrapper.setWidth(width);
    fragmentWrapper.assignEndpoints(assignedEndpoints);
  }

  private static void checkOrThrow(final boolean expr, final Logger logger, final String errMsg, Object... args)
      throws PhysicalOperatorSetupException {
    if (!expr) {
      logger.error(errMsg, args);
      throw new PhysicalOperatorSetupException("Can not parallelize fragment.");
    }
  }
}