lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

Implement Savepoints and required metadata timeline - Part 2

Browse Source
This commit is contained in:
Prasanna Rajaperumal
2017-03-13 23:09:29 -07:00
parent 6f36e1eaaf
commit d83b671ada
7 changed files with 135 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
hoodie-client/src/main/java/com/uber/hoodie/HoodieWriteClient.java
View File

@@ -26,6 +26,7 @@ import com.uber.hoodie.avro.model.HoodieSavepointMetadata;
import com.uber.hoodie.common.HoodieCleanStat;
import com.uber.hoodie.common.HoodieRollbackStat;
import com.uber.hoodie.common.model.HoodieCommitMetadata;
import com.uber.hoodie.common.model.HoodieDataFile;
import com.uber.hoodie.common.model.HoodieKey;
import com.uber.hoodie.common.model.HoodieRecord;
import com.uber.hoodie.common.model.HoodieRecordLocation;
@@ -33,8 +34,10 @@ import com.uber.hoodie.common.model.HoodieRecordPayload;
import com.uber.hoodie.common.model.HoodieWriteStat;
import com.uber.hoodie.common.table.HoodieTableMetaClient;
import com.uber.hoodie.common.table.HoodieTimeline;
import com.uber.hoodie.common.table.TableFileSystemView;
import com.uber.hoodie.common.table.timeline.HoodieActiveTimeline;
import com.uber.hoodie.common.table.timeline.HoodieInstant;
import com.uber.hoodie.common.table.view.HoodieTableFileSystemView;
import com.uber.hoodie.common.util.AvroUtils;
import com.uber.hoodie.common.util.FSUtils;
import com.uber.hoodie.config.HoodieWriteConfig;
@@ -53,9 +56,12 @@ import com.uber.hoodie.metrics.HoodieMetrics;
import com.uber.hoodie.table.HoodieTable;
import com.uber.hoodie.table.WorkloadProfile;
import org.apache.commons.lang3.tuple.Pair;
import org.apache.hadoop.fs.FileStatus;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.LocatedFileStatus;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.RemoteIterator;
import org.apache.log4j.LogManager;
import org.apache.log4j.Logger;
import org.apache.spark.Accumulator;
@@ -63,6 +69,7 @@ import org.apache.spark.Partitioner;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.FlatMapFunction;
import org.apache.spark.api.java.function.Function;
import org.apache.spark.api.java.function.Function2;
import org.apache.spark.api.java.function.PairFunction;
@@ -78,6 +85,7 @@ import java.util.Collections;
import java.util.Date;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
import java.util.Optional;
@@ -396,16 +404,20 @@ public class HoodieWriteClient<T extends HoodieRecordPayload> implements Seriali
}
/**
* Savepoint the latest commit. The data files and commit files for that commit will never be rolledback,
* cleaned or archived. This gives an option to rollback the state to the savepoint anytime.
* Savepoint a specific commit. Latest version of data files as of the passed in commitTime
* will be referenced in the savepoint and will never be cleaned. The savepointed commit
* will never be rolledback or archived.
*
* This gives an option to rollback the state to the savepoint anytime.
* Savepoint needs to be manually created and deleted.
*
* Savepoint should be on a commit that is not cleaned.
* Savepoint should be on a commit that could not have been cleaned.
*
* @param savePointMetadata - metadata about the savepoint
* @param user - User creating the savepoint
* @param comment - Comment for the savepoint
* @return true if the savepoint was created successfully
*/
public boolean savepoint(HoodieSavepointMetadata savePointMetadata) {
public boolean savepoint(String user, String comment) {
HoodieTable<T> table = HoodieTable
.getHoodieTable(new HoodieTableMetaClient(fs, config.getBasePath(), true), config);
if (table.getCompletedCommitTimeline().empty()) {
@@ -414,20 +426,25 @@ public class HoodieWriteClient<T extends HoodieRecordPayload> implements Seriali
String latestCommit = table.getCompletedCommitTimeline().lastInstant().get().getTimestamp();
logger.info("Savepointing latest commit " + latestCommit);
return savepoint(latestCommit, savePointMetadata);
return savepoint(latestCommit, user, comment);
}
/**
* Savepoint a specific commit. The data files and commit files for that commit will never be rolledback,
* cleaned or archived. This gives an option to rollback the state to the savepoint anytime.
* Savepoint a specific commit. Latest version of data files as of the passed in commitTime
* will be referenced in the savepoint and will never be cleaned. The savepointed commit
* will never be rolledback or archived.
*
* This gives an option to rollback the state to the savepoint anytime.
* Savepoint needs to be manually created and deleted.
*
* Savepoint should be on a commit that is not cleaned.
* Savepoint should be on a commit that could not have been cleaned.
*
* @param savePointMetadata - metadata about the savepoint
* @param commitTime - commit that should be savepointed
* @param user - User creating the savepoint
* @param comment - Comment for the savepoint
* @return true if the savepoint was created successfully
*/
public boolean savepoint(String commitTime, HoodieSavepointMetadata savePointMetadata) {
public boolean savepoint(String commitTime, String user, String comment) {
HoodieTable<T> table = HoodieTable
.getHoodieTable(new HoodieTableMetaClient(fs, config.getBasePath(), true), config);
Optional<HoodieInstant> cleanInstant = table.getCompletedCleanTimeline().lastInstant();
@@ -455,10 +472,24 @@ public class HoodieWriteClient<T extends HoodieRecordPayload> implements Seriali
"Could not savepoint commit " + commitTime + " as this is beyond the lookup window "
+ lastCommitRetained);
Map<String, List<String>> latestFilesMap = jsc.parallelize(
FSUtils.getAllPartitionPaths(fs, table.getMetaClient().getBasePath()))
.mapToPair((PairFunction<String, String, List<String>>) partitionPath -> {
// Scan all partitions files with this commit time
logger.info("Collecting latest files in partition path " + partitionPath);
TableFileSystemView view = table.getFileSystemView();
List<String> latestFiles =
view.getLatestVersionInPartition(partitionPath, commitTime)
.map(HoodieDataFile::getFileName).collect(Collectors.toList());
return new Tuple2<String, List<String>>(partitionPath, latestFiles);
}).collectAsMap();
HoodieSavepointMetadata metadata =
AvroUtils.convertSavepointMetadata(user, comment, latestFilesMap);
// Nothing to save in the savepoint
table.getActiveTimeline().saveAsComplete(
new HoodieInstant(false, HoodieTimeline.SAVEPOINT_ACTION, commitTime),
AvroUtils.serializeSavepointMetadata(savePointMetadata));
AvroUtils.serializeSavepointMetadata(metadata));
logger.info("Savepoint " + commitTime + " created");
return true;
} catch (IOException e) {

24
hoodie-client/src/main/java/com/uber/hoodie/io/HoodieCleaner.java
View File

@@ -21,6 +21,7 @@ import com.google.common.collect.Maps;
import com.uber.hoodie.common.HoodieCleanStat;
import com.uber.hoodie.common.model.HoodieCleaningPolicy;
import com.uber.hoodie.common.model.HoodieDataFile;
import com.uber.hoodie.common.model.HoodieRecordPayload;
import com.uber.hoodie.common.table.HoodieTimeline;
import com.uber.hoodie.common.table.TableFileSystemView;
import com.uber.hoodie.common.table.timeline.HoodieInstant;
@@ -32,6 +33,7 @@ import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.log4j.LogManager;
import org.apache.log4j.Logger;
import org.apache.spark.api.java.function.FlatMapFunction;
import java.io.IOException;
import java.util.ArrayList;
@@ -39,7 +41,9 @@ import java.util.Iterator;
import java.util.List;
import java.util.Map;
import java.util.Optional;
import java.util.function.Function;
import java.util.stream.Collectors;
import java.util.stream.Stream;
/**
* Cleaner is responsible for garbage collecting older files in a given partition path, such that
@@ -50,16 +54,16 @@ import java.util.stream.Collectors;
* <p>
* TODO: Should all cleaning be done based on {@link com.uber.hoodie.common.model.HoodieCommitMetadata}
*/
public class HoodieCleaner {
public class HoodieCleaner<T extends HoodieRecordPayload<T>> {
private static Logger logger = LogManager.getLogger(HoodieCleaner.class);
private final TableFileSystemView fileSystemView;
private final HoodieTimeline commitTimeline;
private HoodieTable hoodieTable;
private HoodieTable<T> hoodieTable;
private HoodieWriteConfig config;
private FileSystem fs;
public HoodieCleaner(HoodieTable hoodieTable, HoodieWriteConfig config) {
public HoodieCleaner(HoodieTable<T> hoodieTable, HoodieWriteConfig config) {
this.hoodieTable = hoodieTable;
this.fileSystemView = hoodieTable.getCompactedFileSystemView();
this.commitTimeline = hoodieTable.getCompletedCommitTimeline();
@@ -85,7 +89,9 @@ public class HoodieCleaner {
fileSystemView.getEveryVersionInPartition(partitionPath)
.collect(Collectors.toList());
List<String> deletePaths = new ArrayList<>();
List<String> savepoints = hoodieTable.getSavepoints();
// Collect all the datafiles savepointed by all the savepoints
List<String> savepointedFiles = hoodieTable.getSavepoints().stream()
.flatMap(s -> hoodieTable.getSavepointedDataFiles(s)).collect(Collectors.toList());
for (List<HoodieDataFile> versionsForFileId : fileVersions) {
int keepVersions = config.getCleanerFileVersionsRetained();
@@ -93,8 +99,8 @@ public class HoodieCleaner {
while (commitItr.hasNext() && keepVersions > 0) {
// Skip this most recent version
HoodieDataFile next = commitItr.next();
if(savepoints.contains(next.getCommitTime())) {
// do not clean datafiles that are savepointed
if(savepointedFiles.contains(next.getFileName())) {
// do not clean up a savepoint data file
continue;
}
keepVersions--;
@@ -134,7 +140,9 @@ public class HoodieCleaner {
"Cleaning " + partitionPath + ", retaining latest " + commitsRetained + " commits. ");
List<String> deletePaths = new ArrayList<>();
List<String> savepoints = hoodieTable.getSavepoints();
// Collect all the datafiles savepointed by all the savepoints
List<String> savepointedFiles = hoodieTable.getSavepoints().stream()
.flatMap(s -> hoodieTable.getSavepointedDataFiles(s)).collect(Collectors.toList());
// determine if we have enough commits, to start cleaning.
if (commitTimeline.countInstants() > commitsRetained) {
@@ -152,7 +160,7 @@ public class HoodieCleaner {
// i.e always spare the last commit.
for (HoodieDataFile afile : fileList) {
String fileCommitTime = afile.getCommitTime();
if(savepoints.contains(fileCommitTime)) {
if(savepointedFiles.contains(afile.getFileName())) {
// do not clean up a savepoint data file
continue;
}

37
hoodie-client/src/main/java/com/uber/hoodie/table/HoodieTable.java
View File

@@ -17,12 +17,15 @@
package com.uber.hoodie.table;
import com.google.common.collect.Sets;
import com.uber.hoodie.avro.model.HoodieSavepointMetadata;
import com.uber.hoodie.avro.model.HoodieSavepointPartitionMetadata;
import com.uber.hoodie.common.table.HoodieTableMetaClient;
import com.uber.hoodie.common.table.HoodieTimeline;
import com.uber.hoodie.common.table.TableFileSystemView;
import com.uber.hoodie.common.table.timeline.HoodieActiveTimeline;
import com.uber.hoodie.common.table.timeline.HoodieInstant;
import com.uber.hoodie.common.table.view.HoodieTableFileSystemView;
import com.uber.hoodie.common.util.AvroUtils;
import com.uber.hoodie.config.HoodieWriteConfig;
import com.uber.hoodie.WriteStatus;
import com.uber.hoodie.common.model.HoodieRecord;
@@ -30,13 +33,17 @@ import com.uber.hoodie.common.model.HoodieRecordPayload;
import com.uber.hoodie.exception.HoodieCommitException;
import com.uber.hoodie.exception.HoodieException;
import com.uber.hoodie.exception.HoodieSavepointException;
import org.apache.hadoop.fs.FileSystem;
import org.apache.spark.Partitioner;
import java.io.IOException;
import java.io.Serializable;
import java.util.Iterator;
import java.util.List;
import java.util.function.Function;
import java.util.stream.Collectors;
import java.util.stream.Stream;
/**
* Abstract implementation of a HoodieTable
@@ -138,11 +145,41 @@ public abstract class HoodieTable<T extends HoodieRecordPayload> implements Seri
return getActiveTimeline().getSavePointTimeline().filterCompletedInstants();
}
/**
* Get the list of savepoints in this table
* @return
*/
public List<String> getSavepoints() {
return getCompletedSavepointTimeline().getInstants().map(HoodieInstant::getTimestamp)
.collect(Collectors.toList());
}
/**
* Get the list of data file names savepointed
*
* @param savepointTime
* @return
* @throws IOException
*/
public Stream<String> getSavepointedDataFiles(String savepointTime) {
if (!getSavepoints().contains(savepointTime)) {
throw new HoodieSavepointException(
"Could not get data files for savepoint " + savepointTime + ". No such savepoint.");
}
HoodieInstant instant =
new HoodieInstant(false, HoodieTimeline.SAVEPOINT_ACTION, savepointTime);
HoodieSavepointMetadata metadata = null;
try {
metadata = AvroUtils.deserializeHoodieSavepointMetadata(
getActiveTimeline().getInstantDetails(instant).get());
} catch (IOException e) {
throw new HoodieSavepointException(
"Could not get savepointed data files for savepoint " + savepointTime, e);
}
return metadata.getPartitionMetadata().values().stream()
.flatMap(s -> s.getSavepointDataFile().stream());
}
public HoodieActiveTimeline getActiveTimeline() {
return metaClient.getActiveTimeline();
}