[
https://issues.apache.org/jira/browse/HUDI-1138?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17387767#comment-17387767
]
ASF GitHub Bot commented on HUDI-1138:
--------------------------------------
yihua commented on a change in pull request #3233:
URL: https://github.com/apache/hudi/pull/3233#discussion_r677110738
##########
File path:
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/config/HoodieWriteConfig.java
##########
@@ -228,6 +229,30 @@
+ "files from lake storage, before committing the write. Reduce this
value, if the high number of tasks incur delays for smaller tables "
+ "or low latency writes.");
+ public static final ConfigProperty<String> MARKERS_TYPE = ConfigProperty
Review comment:
Fixed.
##########
File path:
hudi-timeline-service/src/main/java/org/apache/hudi/timeline/service/handlers/MarkerHandler.java
##########
@@ -0,0 +1,484 @@
+/*
+ * 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.hudi.timeline.service.handlers;
+
+import org.apache.hudi.common.config.SerializableConfiguration;
+import org.apache.hudi.common.engine.HoodieEngineContext;
+import org.apache.hudi.common.metrics.Registry;
+import org.apache.hudi.common.model.IOType;
+import org.apache.hudi.common.table.view.FileSystemViewManager;
+import org.apache.hudi.common.util.collection.ImmutablePair;
+import org.apache.hudi.exception.HoodieException;
+import org.apache.hudi.exception.HoodieIOException;
+import org.apache.hudi.timeline.service.TimelineService;
+
+import com.fasterxml.jackson.core.JsonProcessingException;
+import com.fasterxml.jackson.databind.ObjectMapper;
+import io.javalin.Context;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FSDataInputStream;
+import org.apache.hadoop.fs.FSDataOutputStream;
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.util.StringUtils;
+import org.apache.log4j.LogManager;
+import org.apache.log4j.Logger;
+
+import java.io.BufferedReader;
+import java.io.BufferedWriter;
+import java.io.Closeable;
+import java.io.IOException;
+import java.io.InputStreamReader;
+import java.io.OutputStreamWriter;
+import java.nio.charset.StandardCharsets;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Map;
+import java.util.Set;
+import java.util.concurrent.CompletableFuture;
+import java.util.concurrent.Executors;
+import java.util.concurrent.ScheduledExecutorService;
+import java.util.concurrent.TimeUnit;
+import java.util.stream.Collectors;
+
+import static org.apache.hudi.timeline.service.RequestHandler.jsonifyResult;
+
+/**
+ * REST Handler servicing marker requests.
+ *
+ * The marker creation requests are handled asynchronous, while other types of
requests
+ * are handled synchronous.
+ *
+ * Marker creation requests are batch processed periodically by a thread.
Each batch
+ * processing thread adds new markers to a marker file. Given that marker
file operation
+ * can take time, multiple concurrent threads can run at the same, while they
operate
+ * on different marker files storing mutually exclusive marker entries. At
any given
+ * time, a marker file is touched by at most one thread to guarantee
consistency.
+ * Below is an example of running batch processing threads.
+ *
+ * |-----| batch interval
+ * Thread 1 |-------------------------->| writing to MARKERS1
+ * Thread 2 |-------------------------->| writing to MARKERS2
+ * Thread 3 |-------------------------->| writing to MARKERS3
+ */
+public class MarkerHandler extends Handler {
+ public static final String MARKERS_FILENAME_PREFIX = "MARKERS";
+ private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
+ private static final Logger LOG = LogManager.getLogger(MarkerHandler.class);
+ // Margin time for scheduling the processing of the next batch of marker
creation requests
+ private static final long SCHEDULING_MARGIN_TIME_MS = 5L;
+
+ private final Registry metricsRegistry;
+ private final ScheduledExecutorService executorService;
+ // A cached copy of all markers in memory
+ // Mapping: {markerDirPath -> all markers}
+ private final Map<String, Set<String>> allMarkersMap = new HashMap<>();
Review comment:
Yes, that's the intention behind this map. Each writer should write to
an independent marker directory and the markers are maintained based on the
marker directory containing the instant time.
##########
File path:
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/table/marker/TimelineServerBasedMarkerFiles.java
##########
@@ -0,0 +1,178 @@
+/*
+ * 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.hudi.table.marker;
+
+import org.apache.hudi.common.engine.HoodieEngineContext;
+import org.apache.hudi.common.model.IOType;
+import org.apache.hudi.common.util.HoodieTimer;
+import org.apache.hudi.common.util.Option;
+import org.apache.hudi.exception.HoodieRemoteException;
+import org.apache.hudi.table.HoodieTable;
+
+import com.fasterxml.jackson.core.type.TypeReference;
+import com.fasterxml.jackson.databind.ObjectMapper;
+import org.apache.hadoop.fs.Path;
+import org.apache.http.client.fluent.Request;
+import org.apache.http.client.fluent.Response;
+import org.apache.http.client.utils.URIBuilder;
+import org.apache.log4j.LogManager;
+import org.apache.log4j.Logger;
+
+import java.io.IOException;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.Map;
+import java.util.Set;
+import java.util.stream.Collectors;
+
+import static
org.apache.hudi.common.table.marker.MarkerOperation.ALL_MARKERS_URL;
+import static
org.apache.hudi.common.table.marker.MarkerOperation.CREATE_AND_MERGE_MARKERS_URL;
+import static
org.apache.hudi.common.table.marker.MarkerOperation.CREATE_MARKER_URL;
+import static
org.apache.hudi.common.table.marker.MarkerOperation.DELETE_MARKER_DIR_URL;
+import static
org.apache.hudi.common.table.marker.MarkerOperation.MARKERS_DIR_EXISTS_URL;
+import static
org.apache.hudi.common.table.marker.MarkerOperation.MARKER_DIR_PATH_PARAM;
+import static
org.apache.hudi.common.table.marker.MarkerOperation.MARKER_NAME_PARAM;
+
+/**
+ * Marker file operations of using timeline server as a proxy to create and
delete marker files.
+ * Each data file has a corresponding marker entry, which is stored in a
limited number of
+ * marker files maintained by the timeline service (each marker file contains
multiple marker
+ * entries).
+ */
+public class TimelineServerBasedMarkerFiles extends MarkerFiles {
+ private static final Logger LOG =
LogManager.getLogger(TimelineServerBasedMarkerFiles.class);
+ private final ObjectMapper mapper;
+ private final String timelineServerHost;
+ private final int timelineServerPort;
+ private final int timeoutSecs;
+
+ public TimelineServerBasedMarkerFiles(HoodieTable table, String instantTime)
{
+ this(table.getMetaClient().getBasePath(),
+ table.getMetaClient().getMarkerFolderPath(instantTime), instantTime,
+ table.getConfig().getViewStorageConfig().getRemoteViewServerHost(),
+ table.getConfig().getViewStorageConfig().getRemoteViewServerPort(),
+
table.getConfig().getViewStorageConfig().getRemoteTimelineClientTimeoutSecs());
+ }
+
+ TimelineServerBasedMarkerFiles(String basePath, String markerFolderPath,
String instantTime,
+ String timelineServerHost, int
timelineServerPort, int timeoutSecs) {
+ super(basePath, markerFolderPath, instantTime);
+ this.mapper = new ObjectMapper();
+ this.timelineServerHost = timelineServerHost;
+ this.timelineServerPort = timelineServerPort;
+ this.timeoutSecs = timeoutSecs;
+ }
+
+ @Override
+ public boolean deleteMarkerDir(HoodieEngineContext context, int parallelism)
{
+ Map<String, String> paramsMap =
Collections.singletonMap(MARKER_DIR_PATH_PARAM, markerDirPath.toString());
+ try {
+ return executeRequestToTimelineServer(
+ DELETE_MARKER_DIR_URL, paramsMap, new TypeReference<Boolean>() {},
RequestMethod.POST);
+ } catch (IOException e) {
+ throw new HoodieRemoteException(e);
+ }
+ }
+
+ @Override
+ public boolean doesMarkerDirExist() {
+ Map<String, String> paramsMap =
Collections.singletonMap(MARKER_DIR_PATH_PARAM, markerDirPath.toString());
+ try {
+ return executeRequestToTimelineServer(
+ MARKERS_DIR_EXISTS_URL, paramsMap, new TypeReference<Boolean>() {},
RequestMethod.GET);
+ } catch (IOException e) {
+ throw new HoodieRemoteException(e);
Review comment:
Fixed.
##########
File path:
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/table/marker/TimelineServerBasedMarkerFiles.java
##########
@@ -0,0 +1,178 @@
+/*
+ * 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.hudi.table.marker;
+
+import org.apache.hudi.common.engine.HoodieEngineContext;
+import org.apache.hudi.common.model.IOType;
+import org.apache.hudi.common.util.HoodieTimer;
+import org.apache.hudi.common.util.Option;
+import org.apache.hudi.exception.HoodieRemoteException;
+import org.apache.hudi.table.HoodieTable;
+
+import com.fasterxml.jackson.core.type.TypeReference;
+import com.fasterxml.jackson.databind.ObjectMapper;
+import org.apache.hadoop.fs.Path;
+import org.apache.http.client.fluent.Request;
Review comment:
These are used by `RemoteHoodieTableFileSystemView` as well.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
> Re-implement marker files via timeline server
> ---------------------------------------------
>
> Key: HUDI-1138
> URL: https://issues.apache.org/jira/browse/HUDI-1138
> Project: Apache Hudi
> Issue Type: Improvement
> Components: Writer Core
> Affects Versions: 0.9.0
> Reporter: Vinoth Chandar
> Assignee: Ethan Guo
> Priority: Blocker
> Labels: pull-request-available
> Fix For: 0.9.0
>
>
> Even as you can argue that RFC-15/consolidated metadata, removes the need for
> deleting partial files written due to spark task failures/stage retries. It
> will still leave extra files inside the table (and users will pay for it
> every month) and we need the marker mechanism to be able to delete these
> partial files.
> Here we explore if we can improve the current marker file mechanism, that
> creates one marker file per data file written, by
> Delegating the createMarker() call to the driver/timeline server, and have it
> create marker metadata into a single file handle, that is flushed for
> durability guarantees
>
> P.S: I was tempted to think Spark listener mechanism can help us deal with
> failed tasks, but it has no guarantees. the writer job could die without
> deleting a partial file. i.e it can improve things, but cant provide
> guarantees
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
