Skip to content
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.

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

[Flink] Support speculative execution when batch read paimon table. #4395

Merged
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
Original file line number Diff line number Diff line change
@@ -0,0 +1,45 @@
/*
* 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.flink.api.connector.source;

import org.apache.flink.annotation.PublicEvolving;

/**
* An decorative interface of {@link SplitEnumerator} which allows to handle {@link SourceEvent}
* sent from a specific execution attempt.
*
* <p>The split enumerator must implement this interface if it needs to deal with custom source
* events and is used in cases that a subtask can have multiple concurrent execution attempts, e.g.
* if speculative execution is enabled. Otherwise an error will be thrown when the split enumerator
* receives a custom source event.
*/
@PublicEvolving
public interface SupportsHandleExecutionAttemptSourceEvent {

/**
* Handles a custom source event from the source reader. It is similar to {@link
* SplitEnumerator#handleSourceEvent(int, SourceEvent)} but is aware of the subtask execution
* attempt who sent this event.
*
* @param subtaskId the subtask id of the source reader who sent the source event.
* @param attemptNumber the attempt number of the source reader who sent the source event.
* @param sourceEvent the source event from the source reader.
*/
void handleSourceEvent(int subtaskId, int attemptNumber, SourceEvent sourceEvent);
}
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,7 @@
import org.apache.flink.api.connector.source.SplitEnumerator;
import org.apache.flink.api.connector.source.SplitEnumeratorContext;
import org.apache.flink.api.connector.source.SplitsAssignment;
import org.apache.flink.api.connector.source.SupportsHandleExecutionAttemptSourceEvent;
import org.apache.flink.table.connector.source.DynamicFilteringEvent;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
Expand All @@ -41,7 +42,8 @@

/** A {@link SplitEnumerator} implementation for {@link StaticFileStoreSource} input. */
public class StaticFileStoreSplitEnumerator
implements SplitEnumerator<FileStoreSourceSplit, PendingSplitsCheckpoint> {
implements SplitEnumerator<FileStoreSourceSplit, PendingSplitsCheckpoint>,
SupportsHandleExecutionAttemptSourceEvent {
Copy link
Contributor

Choose a reason for hiding this comment

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

You may need to introduce classes for Flink 1.15.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Could we remove 1.15 like remove spark-3.1(#3941)?

We also has plan to support Flink-2.0, to avoid to maintain too many Flink versions, could we remove the support for Flink-1.15?

@JingsongLi

Copy link
Contributor

Choose a reason for hiding this comment

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

No, it is good to keep Flink 1.15 support.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

OK


private static final Logger LOG = LoggerFactory.getLogger(StaticFileStoreSplitEnumerator.class);

Expand Down Expand Up @@ -118,6 +120,17 @@ public Snapshot snapshot() {
return snapshot;
}

@Override
public void handleSourceEvent(int subtaskId, int attemptNumber, SourceEvent sourceEvent) {
// Only recognize events that don't care attemptNumber.
handleSourceEvent(subtaskId, sourceEvent);
}

/**
* When to support a new kind of event, pay attention that whether the new event can be sent
* multiple times from different attempts of one subtask. If so, it should be handled via method
* {@link #handleSourceEvent(int, int, SourceEvent)}
*/
@Override
public void handleSourceEvent(int subtaskId, SourceEvent sourceEvent) {
if (sourceEvent instanceof ReaderConsumeProgressEvent) {

Choose a reason for hiding this comment

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

Maybe add a comment to note that when to support a new kind of event, one needs to pay attention that whether the event can be sent multiple times from different attempts of one subtask. See the description of SupportsHandleExecutionAttemptSourceEvent for more details.

Copy link

@zhuzhurk zhuzhurk Oct 30, 2024

Choose a reason for hiding this comment

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

In case someone later adds a new kind of event directly to this method, ignoring the method with attemptNumber.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

thx, done

Expand Down
Loading