rhauch commented on a change in pull request #11323: URL: https://github.com/apache/kafka/pull/11323#discussion_r733010990
########## File path: connect/runtime/src/main/java/org/apache/kafka/connect/runtime/SubmittedRecords.java ########## @@ -0,0 +1,162 @@ +/* + * 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.kafka.connect.runtime; + +import org.apache.kafka.connect.source.SourceRecord; +import org.apache.kafka.connect.source.SourceTask; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import java.util.Deque; +import java.util.HashMap; +import java.util.LinkedList; +import java.util.Map; + +/** + * Used to track source records that have been (or are about to be) dispatched to a producer and their accompanying + * source offsets. Records are tracked in the order in which they are submitted, which should match the order they were + * returned from {@link SourceTask#poll()}. The latest-eligible offsets for each source partition can be retrieved via + * {@link #committableOffsets()}, where every record up to and including the record for each returned offset has been + * either {@link SubmittedRecord#ack() acknowledged} or {@link #removeLastOccurrence(SubmittedRecord) removed}. + * Note that this class is not thread-safe, though a {@link SubmittedRecord} can be + * {@link SubmittedRecord#ack() acknowledged} from a different thread. + */ +class SubmittedRecords { + + private static final Logger log = LoggerFactory.getLogger(SubmittedRecords.class); + + // Visible for testing + final Map<Map<String, Object>, Deque<SubmittedRecord>> records; Review comment: Great points, @kkonstantine. You're right that using multiple deques does not quite mirror the existing behavior, whereas using a single deque would maintain the same behavior of committing offsets in the same order as they were supplied by the task, albeit more efficiently/effectively than the older behavior. I suspect that most connectors generate messages with proper source partition and source offset maps. However, if any don't then using multiple deques could result in different behavior. IOW, I think the risk is low, but non-zero. Using a single deque avoids that potential issue altogether. Given that this is a bug fix, I agree it's probably better to maintain as much behavior as possible while only fixing the undesirable behavior (in this case, blocking the offset commit thread). @C0urante did such a great job encapsulating the behavior that only the `SubmittedRecords` class would need to be changed/simplified. We can always reevaluate in the future whether to more opportunistically commit offsets for _different_ source partitions independent of the supplied record order (the order they were returned from the task), while committing offsets with the _same_ source partition in the supplied record order. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
