Skip to content

HIVE-27309: Large number of partitions and small files causes OOM in … #4645

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 13, 2023
Merged

HIVE-27309: Large number of partitions and small files causes OOM in … #4645

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 9 additions & 5 deletions ...erg/iceberg-handler/src/main/java/org/apache/iceberg/mr/mapreduce/IcebergInputFormat.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,7 +57,6 @@
import org.apache.iceberg.Scan;
import org.apache.iceberg.Schema;
import org.apache.iceberg.SchemaParser;
import org.apache.iceberg.SerializableTable;
import org.apache.iceberg.SnapshotRef;
import org.apache.iceberg.StructLike;
import org.apache.iceberg.Table;
Expand Down Expand Up @@ -199,7 +198,12 @@ public List<InputSplit> getSplits(JobContext context) {
Configuration conf = context.getConfiguration();
Table table = Optional
.ofNullable(HiveIcebergStorageHandler.table(conf, conf.get(InputFormatConfig.TABLE_IDENTIFIER)))
.orElseGet(() -> Catalogs.loadTable(conf));
.orElseGet(() -> {
Table tbl = Catalogs.loadTable(conf);
conf.set(InputFormatConfig.TABLE_IDENTIFIER, tbl.name());
conf.set(InputFormatConfig.SERIALIZED_TABLE_PREFIX + tbl.name(), SerializationUtil.serializeToBase64(tbl));
return tbl;
});

List<InputSplit> splits = Lists.newArrayList();
boolean applyResidual = !conf.getBoolean(InputFormatConfig.SKIP_RESIDUAL_FILTERING, false);
Expand All @@ -215,14 +219,13 @@ public List<InputSplit> getSplits(JobContext context) {
}

try (CloseableIterable<CombinedScanTask> tasksIterable = scan.planTasks()) {
Table serializableTable = SerializableTable.copyOf(table);
tasksIterable.forEach(task -> {
if (applyResidual && (model == InputFormatConfig.InMemoryDataModel.HIVE ||
model == InputFormatConfig.InMemoryDataModel.PIG)) {
// TODO: We do not support residual evaluation for HIVE and PIG in memory data model yet
checkResiduals(task);
}
splits.add(new IcebergSplit(serializableTable, conf, task));
splits.add(new IcebergSplit(conf, task));
});
} catch (IOException e) {
throw new UncheckedIOException(String.format("Failed to close table scan: %s", scan), e);
Expand Down Expand Up @@ -296,7 +299,8 @@ public void initialize(InputSplit split, TaskAttemptContext newContext) {
CombinedScanTask task = ((IcebergSplit) split).task();
this.context = newContext;
this.conf = newContext.getConfiguration();
this.table = ((IcebergSplit) split).table();
this.table = SerializationUtil.deserializeFromBase64(
conf.get(InputFormatConfig.SERIALIZED_TABLE_PREFIX + conf.get(InputFormatConfig.TABLE_IDENTIFIER)));
Copy link
Contributor

@zhangbutao zhangbutao Sep 4, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
conf.get(InputFormatConfig.SERIALIZED_TABLE_PREFIX + conf.get(InputFormatConfig.TABLE_IDENTIFIER)));
conf.get(InputFormatConfig.SERIALIZED_TABLE_PREFIX + table.name()));

HiveIcebergStorageHandler.checkAndSetIoConfig(conf, table);
this.tasks = task.files().iterator();
this.nameMapping = table.properties().get(TableProperties.DEFAULT_NAME_MAPPING);
Expand Down
17 changes: 1 addition & 16 deletions iceberg/iceberg-handler/src/main/java/org/apache/iceberg/mr/mapreduce/IcebergSplit.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,6 @@
import org.apache.hadoop.mapreduce.InputSplit;
import org.apache.iceberg.CombinedScanTask;
import org.apache.iceberg.FileScanTask;
import org.apache.iceberg.Table;
import org.apache.iceberg.hadoop.Util;
import org.apache.iceberg.mr.InputFormatConfig;
import org.apache.iceberg.util.SerializationUtil;
Expand All @@ -37,7 +36,6 @@ public class IcebergSplit extends InputSplit implements org.apache.hadoop.mapred

public static final String[] ANYWHERE = new String[]{"*"};

private Table table;
private CombinedScanTask task;

private transient String[] locations;
Expand All @@ -47,8 +45,7 @@ public class IcebergSplit extends InputSplit implements org.apache.hadoop.mapred
public IcebergSplit() {
}

IcebergSplit(Table table, Configuration conf, CombinedScanTask task) {
this.table = table;
IcebergSplit(Configuration conf, CombinedScanTask task) {
this.task = task;
this.conf = conf;
}
Expand Down Expand Up @@ -83,27 +80,15 @@ public String[] getLocations() {

@Override
public void write(DataOutput out) throws IOException {
byte[] tableData = SerializationUtil.serializeToBytes(table);
out.writeInt(tableData.length);
out.write(tableData);

byte[] data = SerializationUtil.serializeToBytes(this.task);
out.writeInt(data.length);
out.write(data);
}

@Override
public void readFields(DataInput in) throws IOException {
byte[] tableData = new byte[in.readInt()];
in.readFully(tableData);
this.table = SerializationUtil.deserializeFromBytes(tableData);

byte[] data = new byte[in.readInt()];
in.readFully(data);
this.task = SerializationUtil.deserializeFromBytes(data);
}

public Table table() {
return table;
}
}