Skip to content

KAFKA-9492; Ignore record errors in ProduceResponse for older versions #8030

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
rajinisivaram merged 2 commits into from
Feb 4, 2020
Merged

KAFKA-9492; Ignore record errors in ProduceResponse for older versions #8030

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4e71321
KAFKA-9492; Ignore record errors in ProduceResponse for older versions
rajinisivaram Feb 1, 2020
2585cc3
Address review comment
rajinisivaram Feb 4, 2020
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
24 changes: 13 additions & 11 deletions clients/src/main/java/org/apache/kafka/common/requests/ProduceResponse.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -222,7 +222,8 @@ public ProduceResponse(Struct struct) {
int partition = partRespStruct.get(PARTITION_ID);
Errors error = Errors.forCode(partRespStruct.get(ERROR_CODE));
long offset = partRespStruct.getLong(BASE_OFFSET_KEY_NAME);
long logAppendTime = partRespStruct.getLong(LOG_APPEND_TIME_KEY_NAME);
long logAppendTime = partRespStruct.hasField(LOG_APPEND_TIME_KEY_NAME) ?

@chia7712 chia7712 Feb 3, 2020

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Should we convert the type of LOG_APPEND_TIME_KEY_NAME from long to Field.Int64? the benefit of using Field.Int64 is shown below.

  1. remove the duplicate code of converting LOG_APPEND_TIME_KEY_NAME to Field.Int64
  2. use more readable method - getOrElse

@rajinisivaram rajinisivaram Feb 4, 2020

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

@chia7712 Thanks for the review. As @hachikuji mentioned, we can convert to use the generated classes, so will leave that for a separate PR.

partRespStruct.getLong(LOG_APPEND_TIME_KEY_NAME) : RecordBatch.NO_TIMESTAMP;
long logStartOffset = partRespStruct.getOrElse(LOG_START_OFFSET_FIELD, INVALID_OFFSET);

List<RecordError> recordErrors = Collections.emptyList();
Expand Down Expand Up @@ -274,19 +275,20 @@ protected Struct toStruct(short version) {
partStruct.setIfExists(LOG_APPEND_TIME_KEY_NAME, part.logAppendTime);
partStruct.setIfExists(LOG_START_OFFSET_FIELD, part.logStartOffset);

List<Struct> recordErrors = Collections.emptyList();
if (!part.recordErrors.isEmpty()) {
recordErrors = new ArrayList<>();
for (RecordError indexAndMessage : part.recordErrors) {
Struct indexAndMessageStruct = partStruct.instance(RECORD_ERRORS_KEY_NAME)
.set(BATCH_INDEX_KEY_NAME, indexAndMessage.batchIndex)
.set(BATCH_INDEX_ERROR_MESSAGE_FIELD, indexAndMessage.message);
recordErrors.add(indexAndMessageStruct);
if (partStruct.hasField(RECORD_ERRORS_KEY_NAME)) {
List<Struct> recordErrors = Collections.emptyList();
if (!part.recordErrors.isEmpty()) {
recordErrors = new ArrayList<>();
for (RecordError indexAndMessage : part.recordErrors) {
Struct indexAndMessageStruct = partStruct.instance(RECORD_ERRORS_KEY_NAME)
.set(BATCH_INDEX_KEY_NAME, indexAndMessage.batchIndex)
.set(BATCH_INDEX_ERROR_MESSAGE_FIELD, indexAndMessage.message);
recordErrors.add(indexAndMessageStruct);
}
}
partStruct.set(RECORD_ERRORS_KEY_NAME, recordErrors.toArray());
}

partStruct.setIfExists(RECORD_ERRORS_KEY_NAME, recordErrors.toArray());

partStruct.setIfExists(ERROR_MESSAGE_FIELD, part.errorMessage);
partitionArray.add(partStruct);
}
Expand Down
28 changes: 28 additions & 0 deletions clients/src/test/java/org/apache/kafka/common/requests/RequestResponseTest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -170,6 +170,7 @@

import static java.util.Arrays.asList;
import static org.apache.kafka.common.protocol.ApiKeys.FETCH;
import static org.apache.kafka.common.protocol.ApiKeys.PRODUCE;
import static org.apache.kafka.common.requests.FetchMetadata.INVALID_SESSION_ID;
import static org.apache.kafka.test.TestUtils.toBuffer;
import static org.junit.Assert.assertEquals;
Expand Down Expand Up @@ -653,6 +654,33 @@ public void produceResponseVersionTest() {
assertEquals("Response data does not match", responseData, v2Response.responses());
}

@Test
public void produceResponseRecordErrorsTest() {

@ijuma ijuma Feb 2, 2020

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Maybe this should be in ProduceResponseTest? We started writing this type of test in dedicated classes as this class was getting too large.

@rajinisivaram rajinisivaram Feb 4, 2020

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

@ijuma Thanks for the review. I have moved this test and the other two produce response tests to ProduceResponseTest.

Map<TopicPartition, ProduceResponse.PartitionResponse> responseData = new HashMap<>();
TopicPartition tp = new TopicPartition("test", 0);
ProduceResponse.PartitionResponse partResponse = new ProduceResponse.PartitionResponse(Errors.NONE,
10000, RecordBatch.NO_TIMESTAMP, 100,
Collections.singletonList(new ProduceResponse.RecordError(3, "Record error")),
"Produce failed");
responseData.put(tp, partResponse);

for (short ver = 0; ver <= PRODUCE.latestVersion(); ver++) {
ProduceResponse response = new ProduceResponse(responseData);
Struct struct = response.toStruct(ver);
assertEquals("Should use schema version " + ver, ApiKeys.PRODUCE.responseSchema(ver), struct.schema());
ProduceResponse.PartitionResponse deserialized = new ProduceResponse(struct).responses().get(tp);
if (ver >= 8) {
assertEquals(1, deserialized.recordErrors.size());
assertEquals(3, deserialized.recordErrors.get(0).batchIndex);
assertEquals("Record error", deserialized.recordErrors.get(0).message);
assertEquals("Produce failed", deserialized.errorMessage);
} else {
assertEquals(0, deserialized.recordErrors.size());
assertEquals(null, deserialized.errorMessage);
}
}
}

@Test
public void fetchResponseVersionTest() {
LinkedHashMap<TopicPartition, FetchResponse.PartitionData<MemoryRecords>> responseData = new LinkedHashMap<>();
Expand Down