Skip to content

[CELEBORN-2370] Scope reducer metadata by partition range #3745

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

Open
sunchao wants to merge 2 commits into
base: main
Choose a base branch
from
Open

[CELEBORN-2370] Scope reducer metadata by partition range #3745

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b9a2a29
[CELEBORN-2370] Scope reducer metadata by partition range
sunchao Jun 25, 2026
d16ac36
[CELEBORN-2370] Verify scoped metadata bypasses broadcast
sunchao Jun 25, 2026
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
3 changes: 1 addition & 2 deletions ...park/spark-3/src/main/scala/org/apache/spark/shuffle/celeborn/CelebornShuffleReader.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -166,8 +166,7 @@ class CelebornShuffleReader[K, C](
true
}
try {
// startPartition is irrelevant
fileGroups = shuffleClient.updateFileGroup(shuffleId, startPartition)
fileGroups = shuffleClient.updateFileGroup(shuffleId, startPartition, endPartition)
} catch {
case ce: CelebornIOException
if ce.getCause != null && ce.getCause.isInstanceOf[
Expand Down
6 changes: 6 additions & 0 deletions client/src/main/java/org/apache/celeborn/client/DummyShuffleClient.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -183,6 +183,12 @@ public ShuffleClientImpl.ReduceFileGroups updateFileGroup(int shuffleId, int par
return null;
}

@Override
public ShuffleClientImpl.ReduceFileGroups updateFileGroup(
int shuffleId, int startPartition, int endPartition) throws CelebornIOException {
return null;
}

@Override
public boolean isShuffleStageEnd(int shuffleId) throws Exception {
return true;
Expand Down
31 changes: 31 additions & 0 deletions client/src/main/java/org/apache/celeborn/client/ShuffleClient.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -326,6 +326,37 @@ public abstract void mapPartitionMapperEnd(
public abstract ShuffleClientImpl.ReduceFileGroups updateFileGroup(int shuffleId, int partitionId)
throws CelebornIOException;

public ShuffleClientImpl.ReduceFileGroups updateFileGroup(
int shuffleId, int startPartition, int endPartition) throws CelebornIOException {
if (startPartition < 0 || endPartition < startPartition) {
throw new IllegalArgumentException(
String.format("Invalid reducer file group range [%d, %d)", startPartition, endPartition));
}

ShuffleClientImpl.ReduceFileGroups merged =
new ShuffleClientImpl.ReduceFileGroups(
new ConcurrentHashMap<>(),
null,
ConcurrentHashMap.newKeySet(),
new ConcurrentHashMap<>());
for (int partitionId = startPartition; partitionId < endPartition; partitionId++) {
ShuffleClientImpl.ReduceFileGroups current = updateFileGroup(shuffleId, partitionId);
if (current.partitionGroups != null) {
merged.partitionGroups.putAll(current.partitionGroups);
}
if (current.partitionIds != null) {
merged.partitionIds.addAll(current.partitionIds);
}
if (current.pushFailedBatches != null) {
merged.pushFailedBatches.putAll(current.pushFailedBatches);
}
if (merged.mapAttempts == null) {
merged.mapAttempts = current.mapAttempts;
}
}
return merged;
}

public abstract boolean isShuffleStageEnd(int shuffleId) throws Exception;

// Reduce side read partition which is deduplicated by mapperId+mapperAttemptNum+batchId, batchId
Expand Down
Loading
Loading