lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

[HUDI-2925] Fix duplicate cleaning of same files when unfinished clean operations are present using a config. (#4212)

Browse Source 
Co-authored-by: sivabalan <n.siva.b@gmail.com>
This commit is contained in:
Prashant Wason
2022-02-21 18:53:03 -08:00
committed by GitHub
parent 0c950181aa
commit 0dee8edc97
7 changed files with 118 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

31
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/client/BaseHoodieWriteClient.java
View File

@@ -767,21 +767,28 @@ public abstract class BaseHoodieWriteClient<T extends HoodieRecordPayload, I, K,
if (!tableServicesEnabled(config)) {
return null;
}
if (scheduleInline) {
scheduleTableServiceInternal(cleanInstantTime, Option.empty(), TableServiceType.CLEAN);
}
LOG.info("Cleaner started");
final Timer.Context timerContext = metrics.getCleanCtx();
LOG.info("Cleaned failed attempts if any");
CleanerUtils.rollbackFailedWrites(config.getFailedWritesCleanPolicy(),
HoodieTimeline.CLEAN_ACTION, () -> rollbackFailedWrites(skipLocking));
HoodieCleanMetadata metadata = createTable(config, hadoopConf).clean(context, cleanInstantTime, skipLocking);
if (timerContext != null && metadata != null) {
long durationMs = metrics.getDurationInMs(timerContext.stop());
metrics.updateCleanMetrics(durationMs, metadata.getTotalFilesDeleted());
LOG.info("Cleaned " + metadata.getTotalFilesDeleted() + " files"
+ " Earliest Retained Instant :" + metadata.getEarliestCommitToRetain()
+ " cleanerElapsedMs" + durationMs);
HoodieCleanMetadata metadata = null;
HoodieTable table = createTable(config, hadoopConf);
if (config.allowMultipleCleans() || !table.getActiveTimeline().getCleanerTimeline().filterInflightsAndRequested().firstInstant().isPresent()) {
LOG.info("Cleaner started");
// proceed only if multiple clean schedules are enabled or if there are no pending cleans.
if (scheduleInline) {
scheduleTableServiceInternal(cleanInstantTime, Option.empty(), TableServiceType.CLEAN);
table.getMetaClient().reloadActiveTimeline();
}
metadata = table.clean(context, cleanInstantTime, skipLocking);
if (timerContext != null && metadata != null) {
long durationMs = metrics.getDurationInMs(timerContext.stop());
metrics.updateCleanMetrics(durationMs, metadata.getTotalFilesDeleted());
LOG.info("Cleaned " + metadata.getTotalFilesDeleted() + " files"
+ " Earliest Retained Instant :" + metadata.getEarliestCommitToRetain()
+ " cleanerElapsedMs" + durationMs);
}
}
return metadata;
}

12
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/config/HoodieCompactionConfig.java
View File

@@ -278,6 +278,13 @@ public class HoodieCompactionConfig extends HoodieConfig {
.withDocumentation("The average record size. If not explicitly specified, hudi will compute the "
+ "record size estimate compute dynamically based on commit metadata. "
+ " This is critical in computing the insert parallelism and bin-packing inserts into small files.");
public static final ConfigProperty<Boolean> ALLOW_MULTIPLE_CLEANS = ConfigProperty
.key("hoodie.clean.allow.multiple")
.defaultValue(true)
.sinceVersion("0.11.0")
.withDocumentation("Allows scheduling/executing multiple cleans by enabling this config. If users prefer to strictly ensure clean requests should be mutually exclusive, "
+ ".i.e. a 2nd clean will not be scheduled if another clean is not yet completed to avoid repeat cleaning of same files, they might want to disable this config.");
public static final ConfigProperty<Integer> ARCHIVE_MERGE_FILES_BATCH_SIZE = ConfigProperty
.key("hoodie.archive.merge.files.batch.size")
@@ -642,6 +649,11 @@ public class HoodieCompactionConfig extends HoodieConfig {
return this;
}
public Builder allowMultipleCleans(boolean allowMultipleCleanSchedules) {
compactionConfig.setValue(ALLOW_MULTIPLE_CLEANS, String.valueOf(allowMultipleCleanSchedules));
return this;
}
public Builder withCleanerParallelism(int cleanerParallelism) {
compactionConfig.setValue(CLEANER_PARALLELISM_VALUE, String.valueOf(cleanerParallelism));
return this;

4
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/config/HoodieWriteConfig.java
View File

@@ -1117,6 +1117,10 @@ public class HoodieWriteConfig extends HoodieConfig {
return getInt(HoodieCompactionConfig.COPY_ON_WRITE_RECORD_SIZE_ESTIMATE);
}
public boolean allowMultipleCleans() {
return getBoolean(HoodieCompactionConfig.ALLOW_MULTIPLE_CLEANS);
}
public boolean shouldAutoTuneInsertSplits() {
return getBoolean(HoodieCompactionConfig.COPY_ON_WRITE_AUTO_SPLIT_INSERTS);
}

2
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/table/action/BaseActionExecutor.java
View File

@@ -65,7 +65,7 @@ public abstract class BaseActionExecutor<T extends HoodieRecordPayload, I, K, O,
* Writes clean metadata to table metadata.
* @param metadata clean metadata of interest.
*/
protected final void writeTableMetadata(HoodieCleanMetadata metadata) {
protected final void writeTableMetadata(HoodieCleanMetadata metadata, String instantTime) {
table.getMetadataWriter(instantTime).ifPresent(w -> w.update(metadata, instantTime));
}

8
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/table/action/clean/CleanActionExecutor.java
View File

@@ -208,7 +208,7 @@ public class CleanActionExecutor<T extends HoodieRecordPayload, I, K, O> extends
if (!skipLocking) {
this.txnManager.beginTransaction(Option.empty(), Option.empty());
}
writeTableMetadata(metadata);
writeTableMetadata(metadata, inflightInstant.getTimestamp());
table.getActiveTimeline().transitionCleanInflightToComplete(inflightInstant,
TimelineMetadataUtils.serializeCleanMetadata(metadata));
LOG.info("Marked clean started on " + inflightInstant.getTimestamp() + " as complete");
@@ -240,9 +240,13 @@ public class CleanActionExecutor<T extends HoodieRecordPayload, I, K, O> extends
LOG.warn("Failed to perform previous clean operation, instant: " + hoodieInstant, e);
}
}
table.getMetaClient().reloadActiveTimeline();
if (config.isMetadataTableEnabled()) {
table.getHoodieView().sync();
}
});
table.getMetaClient().reloadActiveTimeline();
}
// return the last clean metadata for now
// TODO (NA) : Clean only the earliest pending clean just like how we do for other table services
// This requires the CleanActionExecutor to be refactored as BaseCommitActionExecutor

68
hudi-client/hudi-spark-client/src/test/java/org/apache/hudi/table/TestCleaner.java
View File

@@ -122,6 +122,7 @@ import static org.apache.hudi.testutils.Assertions.assertNoWriteErrors;
import static org.awaitility.Awaitility.await;
import static org.junit.jupiter.api.Assertions.assertEquals;
import static org.junit.jupiter.api.Assertions.assertFalse;
import static org.junit.jupiter.api.Assertions.assertNotNull;
import static org.junit.jupiter.api.Assertions.assertNull;
import static org.junit.jupiter.api.Assertions.assertTrue;
@@ -254,6 +255,73 @@ public class TestCleaner extends HoodieClientTestBase {
SparkRDDWriteClient::upsertPreppedRecords, true);
}
/**
* Tests no more than 1 clean is scheduled/executed if HoodieCompactionConfig.allowMultipleCleanSchedule config is disabled.
*/
@Test
public void testMultiClean() {
HoodieWriteConfig writeConfig = getConfigBuilder()
.withFileSystemViewConfig(new FileSystemViewStorageConfig.Builder()
.withEnableBackupForRemoteFileSystemView(false).build())
.withCompactionConfig(HoodieCompactionConfig.newBuilder().compactionSmallFileSize(1024 * 1024 * 1024)
.withInlineCompaction(false).withMaxNumDeltaCommitsBeforeCompaction(1)
.withFailedWritesCleaningPolicy(HoodieFailedWritesCleaningPolicy.EAGER)
.allowMultipleCleans(false)
.withAutoClean(false).retainCommits(1).retainFileVersions(1).build())
.withEmbeddedTimelineServerEnabled(false).build();
int index = 0;
String cleanInstantTime;
final String partition = "2015/03/16";
try (SparkRDDWriteClient client = new SparkRDDWriteClient(context, writeConfig)) {
// Three writes so we can initiate a clean
for (; index < 3; ++index) {
String newCommitTime = "00" + index;
List<HoodieRecord> records = dataGen.generateInsertsForPartition(newCommitTime, 1, partition);
client.startCommitWithTime(newCommitTime);
client.insert(jsc.parallelize(records, 1), newCommitTime).collect();
}
}
// mimic failed/leftover clean by scheduling a clean but not performing it
cleanInstantTime = "00" + index++;
HoodieTable table = HoodieSparkTable.create(writeConfig, context);
Option<HoodieCleanerPlan> cleanPlan = table.scheduleCleaning(context, cleanInstantTime, Option.empty());
assertEquals(cleanPlan.get().getFilePathsToBeDeletedPerPartition().get(partition).size(), 1);
assertEquals(metaClient.reloadActiveTimeline().getCleanerTimeline().filterInflightsAndRequested().countInstants(), 1);
try (SparkRDDWriteClient client = new SparkRDDWriteClient(context, writeConfig)) {
// Next commit. This is required so that there is an additional file version to clean.
String newCommitTime = "00" + index++;
List<HoodieRecord> records = dataGen.generateInsertsForPartition(newCommitTime, 1, partition);
client.startCommitWithTime(newCommitTime);
client.insert(jsc.parallelize(records, 1), newCommitTime).collect();
// Initiate another clean. The previous leftover clean will be attempted first, followed by another clean
// due to the commit above.
String newCleanInstantTime = "00" + index++;
HoodieCleanMetadata cleanMetadata = client.clean(newCleanInstantTime);
// subsequent clean should not be triggered since allowMultipleCleanSchedules is set to false
assertNull(cleanMetadata);
// let the old clean complete
table = HoodieSparkTable.create(writeConfig, context);
cleanMetadata = table.clean(context, cleanInstantTime, false);
assertNotNull(cleanMetadata);
// any new clean should go ahead
cleanMetadata = client.clean(newCleanInstantTime);
// subsequent clean should not be triggered since allowMultipleCleanSchedules is set to false
assertNotNull(cleanMetadata);
// 1 file cleaned
assertEquals(cleanMetadata.getPartitionMetadata().get(partition).getSuccessDeleteFiles().size(), 1);
assertEquals(cleanMetadata.getPartitionMetadata().get(partition).getFailedDeleteFiles().size(), 0);
assertEquals(cleanMetadata.getPartitionMetadata().get(partition).getDeletePathPatterns().size(), 1);
}
}
/**
* Test Helper for Cleaning by versions logic from HoodieWriteClient API perspective.
*

8
hudi-common/src/main/java/org/apache/hudi/common/util/CleanerUtils.java
View File

@@ -35,6 +35,9 @@ import org.apache.hudi.common.table.timeline.versioning.clean.CleanMetadataV1Mig
import org.apache.hudi.common.table.timeline.versioning.clean.CleanMetadataV2MigrationHandler;
import org.apache.hudi.common.table.timeline.versioning.clean.CleanPlanMigrator;
import org.apache.log4j.LogManager;
import org.apache.log4j.Logger;
import java.io.IOException;
import java.util.HashMap;
import java.util.List;
@@ -43,6 +46,9 @@ import java.util.Map;
import static org.apache.hudi.common.table.timeline.HoodieTimeline.COMMIT_ACTION;
public class CleanerUtils {
private static final Logger LOG = LogManager.getLogger(CleanerUtils.class);
public static final Integer CLEAN_METADATA_VERSION_1 = CleanMetadataV1MigrationHandler.VERSION;
public static final Integer CLEAN_METADATA_VERSION_2 = CleanMetadataV2MigrationHandler.VERSION;
public static final Integer LATEST_CLEAN_METADATA_VERSION = CLEAN_METADATA_VERSION_2;
@@ -131,6 +137,7 @@ public class CleanerUtils {
// No need to do any special cleanup for failed operations during clean
return;
} else if (cleaningPolicy.isLazy()) {
LOG.info("Cleaned failed attempts if any");
// Perform rollback of failed operations for all types of actions during clean
rollbackFailedWritesFunc.apply();
return;
@@ -140,6 +147,7 @@ public class CleanerUtils {
case COMMIT_ACTION:
// For any other actions, perform rollback of failed writes
if (cleaningPolicy.isEager()) {
LOG.info("Cleaned failed attempts if any");
rollbackFailedWritesFunc.apply();
return;
}