[GitHub] [incubator-druid] clintropolis commented on a change in pull request #8088: Add intermediary data server for shuffle
clintropolis commented on a change in pull request #8088: Add intermediary data
server for shuffle
URL: https://github.com/apache/incubator-druid/pull/8088#discussion_r304684637
##
File path:
indexing-service/src/main/java/org/apache/druid/indexing/worker/IntermediaryDataManager.java
##
@@ -0,0 +1,310 @@
+/*
+ * 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.druid.indexing.worker;
+
+import com.google.common.collect.Iterators;
+import com.google.common.io.Files;
+import com.google.inject.Inject;
+import org.apache.commons.io.FileUtils;
+import org.apache.commons.lang3.mutable.MutableInt;
+import org.apache.druid.client.indexing.IndexingServiceClient;
+import org.apache.druid.client.indexing.TaskStatus;
+import org.apache.druid.guice.ManageLifecycle;
+import org.apache.druid.indexing.common.config.TaskConfig;
+import org.apache.druid.indexing.worker.config.WorkerConfig;
+import org.apache.druid.java.util.common.DateTimes;
+import org.apache.druid.java.util.common.IOE;
+import org.apache.druid.java.util.common.ISE;
+import org.apache.druid.java.util.common.concurrent.Execs;
+import org.apache.druid.java.util.common.lifecycle.LifecycleStart;
+import org.apache.druid.java.util.common.lifecycle.LifecycleStop;
+import org.apache.druid.java.util.common.logger.Logger;
+import org.apache.druid.segment.loading.StorageLocation;
+import org.apache.druid.timeline.DataSegment;
+import org.joda.time.DateTime;
+import org.joda.time.Interval;
+import org.joda.time.Period;
+
+import javax.annotation.Nullable;
+import java.io.File;
+import java.io.IOException;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Map;
+import java.util.Map.Entry;
+import java.util.Set;
+import java.util.concurrent.ConcurrentHashMap;
+import java.util.concurrent.ScheduledExecutorService;
+import java.util.concurrent.TimeUnit;
+import java.util.stream.Collectors;
+
+/**
+ * This class manages intermediary segments for data shuffle between native
parallel index tasks.
+ * In native parallel indexing, phase 1 tasks store segment files in local
storage of middleManagers
+ * and phase 2 tasks read those files via HTTP.
+ *
+ * The directory where segment files are placed is structured as
+ * {@link
StorageLocation#path}/supervisorTaskId/startTimeOfSegment/endTimeOfSegment/partitionIdOfSegment.
+ *
+ * This class provides interfaces to store, find, and remove segment files.
+ * It also has a self-cleanup mechanism to clean up stale segment files. It
periodically checks the last access time
+ * per supervisorTask and removes its all segment files if the supervisorTask
is not running anymore.
+ */
+@ManageLifecycle
+public class IntermediaryDataManager
+{
+ private static final Logger log = new Logger(IntermediaryDataManager.class);
+
+ private final long intermediaryPartitionDiscoveryPeriodSec;
+ private final long intermediaryPartitionCleanupPeriodSec;
+ private final Period intermediaryPartitionTimeout;
+ private final List shuffleDataLocations;
+ private final IndexingServiceClient indexingServiceClient;
+
+ // supervisorTaskId -> time to check supervisorTask status
+ // This time is initialized when a new supervisorTask is found and updated
whenever a partition is accessed for
+ // the supervisor.
+ private final ConcurrentHashMap supervisorTaskCheckTimes =
new ConcurrentHashMap<>();
+
+ // supervisorTaskId -> cyclic iterator of storage locations
+ private final Map> locationIterators = new
HashMap<>();
+
+ // The overlord is supposed to send a cleanup request as soon as the
supervisorTask is finished in parallel indexing,
+ // but middleManager or indexer could miss the request. This executor is to
automatically clean up unused intermediary
+ // partitions.
+ // This can be null until IntermediaryDataManager is started.
+ @Nullable
+ private ScheduledExecutorService supervisorTaskChecker;
+
+ @Inject
+ public IntermediaryDataManager(
+ WorkerConfig workerConfig,
+ TaskConfig taskConfig,
+ IndexingServiceClient indexingServiceClient
+ )
+ {
+this.intermediaryPartitio
[GitHub] [incubator-druid] clintropolis commented on a change in pull request #8088: Add intermediary data server for shuffle
clintropolis commented on a change in pull request #8088: Add intermediary data
server for shuffle
URL: https://github.com/apache/incubator-druid/pull/8088#discussion_r304684720
##
File path:
indexing-service/src/main/java/org/apache/druid/indexing/worker/IntermediaryDataManager.java
##
@@ -0,0 +1,310 @@
+/*
+ * 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.druid.indexing.worker;
+
+import com.google.common.collect.Iterators;
+import com.google.common.io.Files;
+import com.google.inject.Inject;
+import org.apache.commons.io.FileUtils;
+import org.apache.commons.lang3.mutable.MutableInt;
+import org.apache.druid.client.indexing.IndexingServiceClient;
+import org.apache.druid.client.indexing.TaskStatus;
+import org.apache.druid.guice.ManageLifecycle;
+import org.apache.druid.indexing.common.config.TaskConfig;
+import org.apache.druid.indexing.worker.config.WorkerConfig;
+import org.apache.druid.java.util.common.DateTimes;
+import org.apache.druid.java.util.common.IOE;
+import org.apache.druid.java.util.common.ISE;
+import org.apache.druid.java.util.common.concurrent.Execs;
+import org.apache.druid.java.util.common.lifecycle.LifecycleStart;
+import org.apache.druid.java.util.common.lifecycle.LifecycleStop;
+import org.apache.druid.java.util.common.logger.Logger;
+import org.apache.druid.segment.loading.StorageLocation;
+import org.apache.druid.timeline.DataSegment;
+import org.joda.time.DateTime;
+import org.joda.time.Interval;
+import org.joda.time.Period;
+
+import javax.annotation.Nullable;
+import java.io.File;
+import java.io.IOException;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Map;
+import java.util.Map.Entry;
+import java.util.Set;
+import java.util.concurrent.ConcurrentHashMap;
+import java.util.concurrent.ScheduledExecutorService;
+import java.util.concurrent.TimeUnit;
+import java.util.stream.Collectors;
+
+/**
+ * This class manages intermediary segments for data shuffle between native
parallel index tasks.
+ * In native parallel indexing, phase 1 tasks store segment files in local
storage of middleManagers
+ * and phase 2 tasks read those files via HTTP.
+ *
+ * The directory where segment files are placed is structured as
+ * {@link
StorageLocation#path}/supervisorTaskId/startTimeOfSegment/endTimeOfSegment/partitionIdOfSegment.
+ *
+ * This class provides interfaces to store, find, and remove segment files.
+ * It also has a self-cleanup mechanism to clean up stale segment files. It
periodically checks the last access time
+ * per supervisorTask and removes its all segment files if the supervisorTask
is not running anymore.
+ */
+@ManageLifecycle
+public class IntermediaryDataManager
+{
+ private static final Logger log = new Logger(IntermediaryDataManager.class);
+
+ private final long intermediaryPartitionDiscoveryPeriodSec;
+ private final long intermediaryPartitionCleanupPeriodSec;
+ private final Period intermediaryPartitionTimeout;
+ private final List shuffleDataLocations;
+ private final IndexingServiceClient indexingServiceClient;
+
+ // supervisorTaskId -> time to check supervisorTask status
+ // This time is initialized when a new supervisorTask is found and updated
whenever a partition is accessed for
+ // the supervisor.
+ private final ConcurrentHashMap supervisorTaskCheckTimes =
new ConcurrentHashMap<>();
+
+ // supervisorTaskId -> cyclic iterator of storage locations
+ private final Map> locationIterators = new
HashMap<>();
+
+ // The overlord is supposed to send a cleanup request as soon as the
supervisorTask is finished in parallel indexing,
+ // but middleManager or indexer could miss the request. This executor is to
automatically clean up unused intermediary
+ // partitions.
+ // This can be null until IntermediaryDataManager is started.
+ @Nullable
+ private ScheduledExecutorService supervisorTaskChecker;
+
+ @Inject
+ public IntermediaryDataManager(
+ WorkerConfig workerConfig,
+ TaskConfig taskConfig,
+ IndexingServiceClient indexingServiceClient
+ )
+ {
+this.intermediaryPartitio
[GitHub] [incubator-druid] clintropolis commented on a change in pull request #8088: Add intermediary data server for shuffle
clintropolis commented on a change in pull request #8088: Add intermediary data
server for shuffle
URL: https://github.com/apache/incubator-druid/pull/8088#discussion_r304684892
##
File path:
indexing-service/src/main/java/org/apache/druid/indexing/worker/IntermediaryDataManager.java
##
@@ -0,0 +1,310 @@
+/*
+ * 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.druid.indexing.worker;
+
+import com.google.common.collect.Iterators;
+import com.google.common.io.Files;
+import com.google.inject.Inject;
+import org.apache.commons.io.FileUtils;
+import org.apache.commons.lang3.mutable.MutableInt;
+import org.apache.druid.client.indexing.IndexingServiceClient;
+import org.apache.druid.client.indexing.TaskStatus;
+import org.apache.druid.guice.ManageLifecycle;
+import org.apache.druid.indexing.common.config.TaskConfig;
+import org.apache.druid.indexing.worker.config.WorkerConfig;
+import org.apache.druid.java.util.common.DateTimes;
+import org.apache.druid.java.util.common.IOE;
+import org.apache.druid.java.util.common.ISE;
+import org.apache.druid.java.util.common.concurrent.Execs;
+import org.apache.druid.java.util.common.lifecycle.LifecycleStart;
+import org.apache.druid.java.util.common.lifecycle.LifecycleStop;
+import org.apache.druid.java.util.common.logger.Logger;
+import org.apache.druid.segment.loading.StorageLocation;
+import org.apache.druid.timeline.DataSegment;
+import org.joda.time.DateTime;
+import org.joda.time.Interval;
+import org.joda.time.Period;
+
+import javax.annotation.Nullable;
+import java.io.File;
+import java.io.IOException;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Map;
+import java.util.Map.Entry;
+import java.util.Set;
+import java.util.concurrent.ConcurrentHashMap;
+import java.util.concurrent.ScheduledExecutorService;
+import java.util.concurrent.TimeUnit;
+import java.util.stream.Collectors;
+
+/**
+ * This class manages intermediary segments for data shuffle between native
parallel index tasks.
+ * In native parallel indexing, phase 1 tasks store segment files in local
storage of middleManagers
+ * and phase 2 tasks read those files via HTTP.
+ *
+ * The directory where segment files are placed is structured as
+ * {@link
StorageLocation#path}/supervisorTaskId/startTimeOfSegment/endTimeOfSegment/partitionIdOfSegment.
+ *
+ * This class provides interfaces to store, find, and remove segment files.
+ * It also has a self-cleanup mechanism to clean up stale segment files. It
periodically checks the last access time
+ * per supervisorTask and removes its all segment files if the supervisorTask
is not running anymore.
+ */
+@ManageLifecycle
+public class IntermediaryDataManager
+{
+ private static final Logger log = new Logger(IntermediaryDataManager.class);
+
+ private final long intermediaryPartitionDiscoveryPeriodSec;
+ private final long intermediaryPartitionCleanupPeriodSec;
+ private final Period intermediaryPartitionTimeout;
+ private final List shuffleDataLocations;
+ private final IndexingServiceClient indexingServiceClient;
+
+ // supervisorTaskId -> time to check supervisorTask status
+ // This time is initialized when a new supervisorTask is found and updated
whenever a partition is accessed for
+ // the supervisor.
+ private final ConcurrentHashMap supervisorTaskCheckTimes =
new ConcurrentHashMap<>();
+
+ // supervisorTaskId -> cyclic iterator of storage locations
+ private final Map> locationIterators = new
HashMap<>();
+
+ // The overlord is supposed to send a cleanup request as soon as the
supervisorTask is finished in parallel indexing,
+ // but middleManager or indexer could miss the request. This executor is to
automatically clean up unused intermediary
+ // partitions.
+ // This can be null until IntermediaryDataManager is started.
+ @Nullable
+ private ScheduledExecutorService supervisorTaskChecker;
+
+ @Inject
+ public IntermediaryDataManager(
+ WorkerConfig workerConfig,
+ TaskConfig taskConfig,
+ IndexingServiceClient indexingServiceClient
+ )
+ {
+this.intermediaryPartitio
[GitHub] [incubator-druid] clintropolis commented on a change in pull request #8088: Add intermediary data server for shuffle
clintropolis commented on a change in pull request #8088: Add intermediary data
server for shuffle
URL: https://github.com/apache/incubator-druid/pull/8088#discussion_r304674327
##
File path:
indexing-service/src/main/java/org/apache/druid/indexing/worker/http/ShuffleResource.java
##
@@ -0,0 +1,122 @@
+/*
+ * 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.druid.indexing.worker.http;
+
+import com.google.common.io.ByteStreams;
+import com.google.inject.Inject;
+import com.sun.jersey.spi.container.ResourceFilters;
+import org.apache.druid.indexing.worker.IntermediaryDataManager;
+import org.apache.druid.java.util.common.DateTimes;
+import org.apache.druid.java.util.common.StringUtils;
+import org.apache.druid.java.util.common.logger.Logger;
+import org.apache.druid.server.http.security.StateResourceFilter;
+import org.joda.time.Interval;
+
+import javax.ws.rs.DELETE;
+import javax.ws.rs.GET;
+import javax.ws.rs.Path;
+import javax.ws.rs.PathParam;
+import javax.ws.rs.Produces;
+import javax.ws.rs.QueryParam;
+import javax.ws.rs.core.MediaType;
+import javax.ws.rs.core.Response;
+import javax.ws.rs.core.Response.Status;
+import javax.ws.rs.core.StreamingOutput;
+import java.io.File;
+import java.io.FileInputStream;
+import java.io.IOException;
+import java.util.List;
+
+/**
+ * HTTP endpoints for shuffle system. The MiddleManager and Indexer use this
resource to serve intermediary shuffle
+ * data.
+ *
+ * We use {@link StateResourceFilter} here because it performs an admin-like
authorization and
+ * all endpoints here are supposed to be used for only internal communcation.
+ * Another possible alternate could be performing datasource-level
authorization as in TaskResourceFilter.
+ * However, datasource information is not available in middleManagers or
indexers yet which makes hard to use it.
+ * We could develop a new ResourceFileter in the future if needed.
Review comment:
:+1:
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:
us...@infra.apache.org
With regards,
Apache Git Services
-
To unsubscribe, e-mail: commits-unsubscr...@druid.apache.org
For additional commands, e-mail: commits-h...@druid.apache.org
[GitHub] [incubator-druid] clintropolis commented on a change in pull request #8088: Add intermediary data server for shuffle
clintropolis commented on a change in pull request #8088: Add intermediary data server for shuffle URL: https://github.com/apache/incubator-druid/pull/8088#discussion_r304204101 ## File path: server/src/main/java/org/apache/druid/server/http/security/TaskShuffleResourceFilter.java ## @@ -0,0 +1,72 @@ +/* + * 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.druid.server.http.security; + +import com.google.common.base.Preconditions; +import com.google.inject.Inject; +import com.sun.jersey.spi.container.ContainerRequest; +import org.apache.druid.server.security.Access; +import org.apache.druid.server.security.AuthorizationUtils; +import org.apache.druid.server.security.AuthorizerMapper; +import org.apache.druid.server.security.ForbiddenException; +import org.apache.druid.server.security.Resource; +import org.apache.druid.server.security.ResourceAction; +import org.apache.druid.server.security.ResourceType; + +/** + * This resource filter is used for data shuffle between native parallel index tasks. See ShuffleResource for details. + * + * It currently performs the authorization check for DATASOURCE, but ideally, it should be the authorization check + * for task data. This issue should be addressed in the future. Review comment: Could you clarify in these javadocs that I think you are referring to this should be doing what `TaskResourceFilter` is doing? 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: us...@infra.apache.org With regards, Apache Git Services - To unsubscribe, e-mail: commits-unsubscr...@druid.apache.org For additional commands, e-mail: commits-h...@druid.apache.org
[GitHub] [incubator-druid] clintropolis commented on a change in pull request #8088: Add intermediary data server for shuffle
clintropolis commented on a change in pull request #8088: Add intermediary data
server for shuffle
URL: https://github.com/apache/incubator-druid/pull/8088#discussion_r304182914
##
File path:
indexing-service/src/test/java/org/apache/druid/indexing/worker/IntermediaryDataManagerManualAddAndDeleteTest.java
##
@@ -0,0 +1,147 @@
+/*
+ * 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.druid.indexing.worker;
+
+import com.amazonaws.util.StringUtils;
+import com.google.common.collect.ImmutableList;
+import org.apache.commons.io.FileUtils;
+import org.apache.druid.client.indexing.IndexingServiceClient;
+import org.apache.druid.client.indexing.NoopIndexingServiceClient;
+import org.apache.druid.indexing.common.config.TaskConfig;
+import org.apache.druid.indexing.worker.config.WorkerConfig;
+import org.apache.druid.java.util.common.Intervals;
+import org.apache.druid.segment.loading.StorageLocationConfig;
+import org.apache.druid.timeline.DataSegment;
+import org.apache.druid.timeline.partition.NumberedShardSpec;
+import org.joda.time.Interval;
+import org.junit.After;
+import org.junit.Assert;
+import org.junit.Before;
+import org.junit.Rule;
+import org.junit.Test;
+import org.junit.rules.TemporaryFolder;
+
+import java.io.File;
+import java.io.IOException;
+import java.util.Comparator;
+import java.util.List;
+
+public class IntermediaryDataManagerManualAddAndDeleteTest
+{
+ @Rule
+ public TemporaryFolder tempDir = new TemporaryFolder();
+
+ private IntermediaryDataManager intermediaryDataManager;
+
+ @Before
+ public void setup() throws IOException
+ {
+final WorkerConfig workerConfig = new WorkerConfig();
+final TaskConfig taskConfig = new TaskConfig(
+null,
+null,
+null,
+null,
+null,
+false,
+null,
+null,
+ImmutableList.of(new StorageLocationConfig(tempDir.newFolder(), null,
null))
+);
+final IndexingServiceClient indexingServiceClient = new
NoopIndexingServiceClient();
+intermediaryDataManager = new IntermediaryDataManager(workerConfig,
taskConfig, indexingServiceClient);
+intermediaryDataManager.start();
+ }
+
+ @After
+ public void teardown() throws InterruptedException
+ {
+intermediaryDataManager.stop();
+ }
+
+ @Test
+ public void testAddSegment() throws IOException
Review comment:
nit: is this test worth having since it is also implicitly tested by the
other 2 tests?
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:
us...@infra.apache.org
With regards,
Apache Git Services
-
To unsubscribe, e-mail: commits-unsubscr...@druid.apache.org
For additional commands, e-mail: commits-h...@druid.apache.org
