apex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From chandnisingh <...@git.apache.org>
Subject [GitHub] incubator-apex-malhar pull request: APEXMALHAR-1897 added managed ...
Date Wed, 23 Mar 2016 21:33:49 GMT
Github user chandnisingh commented on a diff in the pull request:

    https://github.com/apache/incubator-apex-malhar/pull/145#discussion_r57241107
  
    --- Diff: library/src/main/java/com/datatorrent/lib/state/managed/AbstractManagedStateImpl.java
---
    @@ -0,0 +1,499 @@
    +/**
    + * 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 com.datatorrent.lib.state.managed;
    +
    +import java.io.IOException;
    +import java.util.Comparator;
    +import java.util.Map;
    +import java.util.concurrent.Callable;
    +import java.util.concurrent.ExecutorService;
    +import java.util.concurrent.Executors;
    +import java.util.concurrent.atomic.AtomicReference;
    +
    +import javax.validation.constraints.Min;
    +import javax.validation.constraints.NotNull;
    +
    +import org.joda.time.Duration;
    +import org.slf4j.Logger;
    +import org.slf4j.LoggerFactory;
    +
    +import com.esotericsoftware.kryo.serializers.FieldSerializer;
    +import com.esotericsoftware.kryo.serializers.JavaSerializer;
    +import com.google.common.annotations.VisibleForTesting;
    +import com.google.common.base.Preconditions;
    +import com.google.common.base.Throwables;
    +import com.google.common.collect.Maps;
    +
    +import com.datatorrent.api.Component;
    +import com.datatorrent.api.Context.DAGContext;
    +import com.datatorrent.api.Context.OperatorContext;
    +import com.datatorrent.api.Operator;
    +import com.datatorrent.api.annotation.Stateless;
    +import com.datatorrent.common.util.NameableThreadFactory;
    +import com.datatorrent.lib.fileaccess.FileAccess;
    +import com.datatorrent.lib.fileaccess.TFileImpl;
    +import com.datatorrent.lib.util.comparator.SliceComparator;
    +import com.datatorrent.netlet.util.Slice;
    +
    +/**
    + * An abstract implementation of managed state.<br/>
    + *
    + * The important sub-components here are:
    + * <ol>
    + *   <li>
    + *     {@link #checkpointManager}: writes incremental checkpoints in window files and
transfers data from window
    + *     files to bucket files.
    + *   </li>
    + *   <li>
    + *     {@link #bucketsFileSystem}: a bucket on disk is sub-divided into time-buckets.
This manages meta-bucket
    + *     information (list of {@link BucketsFileSystem.TimeBucketMeta}) per bucket.
    + *   </li>
    + *   <li>
    + *     {@link #timeBucketAssigner}: assigns time-buckets to keys and manages the time
boundaries.
    + *   </li>
    + *   <li>
    + *     {@link #stateTracker}: tracks the size of data in memory and requests buckets
to free memory when enough memory
    + *     is not available.
    + *   </li>
    + *   <li>
    + *     {@link #fileAccess}: plug-able file system abstraction.
    + *   </li>
    + * </ol>
    + *
    + * The implementations of put, getSync and getAsync here use windowId as the time field
to derive timeBucket of a key.
    + */
    +public abstract class AbstractManagedStateImpl
    +    implements ManagedState, Component<OperatorContext>, Operator.CheckpointNotificationListener,
ManagedStateContext,
    +    TimeBucketAssigner.PurgeListener
    +{
    +  private long maxMemorySize;
    +
    +  protected int numBuckets;
    +
    +  @NotNull
    +  private FileAccess fileAccess = new TFileImpl.DTFileImpl();
    +  @NotNull
    +  protected TimeBucketAssigner timeBucketAssigner = new TimeBucketAssigner();
    +
    +  protected Bucket[] buckets;
    +
    +  @Min(1)
    +  private int numReaders = 1;
    +  @NotNull
    +  protected transient ExecutorService readerService;
    +
    +  @NotNull
    +  protected IncrementalCheckpointManager checkpointManager = new IncrementalCheckpointManager();
    +
    +  @NotNull
    +  protected BucketsFileSystem bucketsFileSystem = new BucketsFileSystem();
    +
    +  protected transient OperatorContext operatorContext;
    +  protected transient long windowId;
    +
    +  @NotNull
    +  protected Comparator<Slice> keyComparator = new SliceComparator();
    +
    +  protected final transient AtomicReference<Throwable> throwable = new AtomicReference<>();
    +
    +  @NotNull
    +  @FieldSerializer.Bind(JavaSerializer.class)
    +  private Duration checkStateSizeInterval = Duration.millis(
    +      DAGContext.STREAMING_WINDOW_SIZE_MILLIS.defaultValue * OperatorContext.APPLICATION_WINDOW_COUNT.defaultValue);
    +
    +  @FieldSerializer.Bind(JavaSerializer.class)
    +  private Duration durationPreventingFreeingSpace;
    +
    +  private transient StateTracker stateTracker = new StateTracker();
    +
    +  //accessible to StateTracker
    +  final transient Object commitLock = new Object();
    +
    +  @Override
    +  public void setup(OperatorContext context)
    +  {
    +    operatorContext = context;
    +    fileAccess.init();
    +
    +    timeBucketAssigner.setPurgeListener(this);
    +
    +    //setup all the managed state components
    +    timeBucketAssigner.setup(this);
    +    checkpointManager.setup(this);
    +    bucketsFileSystem.setup(this);
    +
    +    if (buckets == null) {
    +      //create buckets array only once at start when it is not created.
    +      numBuckets = getNumBuckets();
    +      buckets = new Bucket[numBuckets];
    +    }
    +    for (Bucket bucket : buckets) {
    +      if (bucket != null) {
    --- End diff --
    
    Yes it is possible. Let's say user specified numBuckets = 4. This will create ```buckets```
array of size 4. However if we don't see a key for bucket 3 then ```buckets[2]``` will never
be initialized to a a bucket. We are following a lazy initialization strategy.


---
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 infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

Mime
View raw message