hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From GitBox <...@apache.org>
Subject [GitHub] [hadoop] nandakumar131 commented on a change in pull request #620: HDDS-1205. Refactor ReplicationManager to handle QUASI_CLOSED contain…
Date Tue, 19 Mar 2019 06:21:47 GMT
nandakumar131 commented on a change in pull request #620: HDDS-1205. Refactor ReplicationManager
to handle QUASI_CLOSED contain…
URL: https://github.com/apache/hadoop/pull/620#discussion_r266744423
 
 

 ##########
 File path: hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/container/ReplicationManager.java
 ##########
 @@ -0,0 +1,686 @@
+/**
+ * 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
+ * <p>
+ * http://www.apache.org/licenses/LICENSE-2.0
+ * <p>
+ * 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.hadoop.hdds.scm.container;
+
+import com.google.common.annotations.VisibleForTesting;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.hdds.protocol.DatanodeDetails;
+import org.apache.hadoop.hdds.protocol.proto.HddsProtos.LifeCycleState;
+import org.apache.hadoop.hdds.protocol.proto
+    .StorageContainerDatanodeProtocolProtos.ContainerReplicaProto.State;
+import org.apache.hadoop.hdds.scm.ScmConfigKeys;
+import org.apache.hadoop.hdds.scm.container.placement.algorithms
+    .ContainerPlacementPolicy;
+import org.apache.hadoop.hdds.scm.events.SCMEvents;
+import org.apache.hadoop.hdds.server.events.EventPublisher;
+import org.apache.hadoop.ozone.lock.LockManager;
+import org.apache.hadoop.ozone.protocol.commands.CloseContainerCommand;
+import org.apache.hadoop.ozone.protocol.commands.CommandForDatanode;
+import org.apache.hadoop.ozone.protocol.commands.DeleteContainerCommand;
+import org.apache.hadoop.ozone.protocol.commands.ReplicateContainerCommand;
+import org.apache.hadoop.util.ExitUtil;
+import org.apache.hadoop.util.Time;
+import org.apache.ratis.util.Preconditions;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.Iterator;
+import java.util.LinkedHashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.Set;
+import java.util.UUID;
+import java.util.concurrent.TimeUnit;
+import java.util.function.Predicate;
+import java.util.stream.Collectors;
+
+/**
+ * Replication Manager (RM) is the one which is responsible for making sure
+ * that the containers are properly replicated. Replication Manager deals only
+ * with Quasi Closed / Closed container.
+ */
+public class ReplicationManager {
+
+  private static final Logger LOG =
+      LoggerFactory.getLogger(ReplicationManager.class);
+
+  /**
+   * Reference to the ContainerManager.
+   */
+  private final ContainerManager containerManager;
+
+  /**
+   * PlacementPolicy which is used to identify where a container
+   * should be copied.
+   */
+  private final ContainerPlacementPolicy containerPlacement;
+
+  /**
+   * EventPublisher to fire Replicate and Delete container commands.
+   */
+  private final EventPublisher eventPublisher;
+
+  /**
+   * Used for locking a container with its ID while processing it.
+   */
+  private final LockManager<ContainerID> lockManager;
 
 Review comment:
   Sure, we can do it.
   The reason for not having the lock embedded inside the container object is because the
ReplicationManager never changes the state of the container. We get the snapshot of container
state from ContainerManager and take actions, but we never update or modify the container.
   
   Ideally, we don't even need a lock here if we change the logic of `processContainersNow`
method and just interrupt the thread. I was being a bit overcautious and introduced the lock
:)
   
   Modification done by report processors to container state will be protected by the lock
inside ContainerManager.

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org


Mime
View raw message