[
https://issues.apache.org/jira/browse/FLINK-6004?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16346751#comment-16346751
]
ASF GitHub Bot commented on FLINK-6004:
---------------------------------------
Github user tzulitai commented on a diff in the pull request:
https://github.com/apache/flink/pull/5269#discussion_r165039352
--- Diff:
flink-connectors/flink-connector-kinesis/src/main/java/org/apache/flink/streaming/connectors/kinesis/internals/KinesisDataFetcher.java
---
@@ -484,7 +484,10 @@ protected Properties getConsumerConfiguration() {
*/
protected final void emitRecordAndUpdateState(T record, long
recordTimestamp, int shardStateIndex, SequenceNumber lastSequenceNumber) {
synchronized (checkpointLock) {
- sourceContext.collectWithTimestamp(record,
recordTimestamp);
+ if (record != null) {
+ sourceContext.collectWithTimestamp(record,
recordTimestamp);
--- End diff --
We currently do not have a log for that.
I'll add a warning log if the record is null.
> Allow FlinkKinesisConsumer to skip corrupted messages
> -----------------------------------------------------
>
> Key: FLINK-6004
> URL: https://issues.apache.org/jira/browse/FLINK-6004
> Project: Flink
> Issue Type: Improvement
> Components: Streaming Connectors
> Reporter: Tzu-Li (Gordon) Tai
> Assignee: Tzu-Li (Gordon) Tai
> Priority: Major
>
> It is quite clear from the fix of FLINK-3679 that in reality, users might
> encounter corrupted messages from Kafka / Kinesis / generally external
> sources when deserializing them.
> The consumers should support simply skipping those messages, by letting the
> deserialization schema return {{null}}, and checking {{null}} values within
> the consumer.
> This has been done for the Kafka consumer already. This ticket tracks the
> improvement for the Kinesis consumer.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
