AHeise commented on a change in pull request #17189:
URL: https://github.com/apache/flink/pull/17189#discussion_r705021281
##########
File path:
flink-connectors/flink-connector-kinesis/src/main/java/org/apache/flink/streaming/connectors/kinesis/internals/KinesisDataFetcher.java
##########
@@ -819,7 +819,7 @@ public void shutdownFetcher() {
LOG.warn("Encountered exception closing record publisher
factory", e);
}
} finally {
- shardConsumersExecutor.shutdownNow();
+ shardConsumersExecutor.shutdown();
Review comment:
The whole point of https://issues.apache.org/jira/browse/FLINK-23528 is
to get rid of interruptions and thus avoid inconsistent state during
stop-with-savepoint.
Instead of interruptions, the poll loop should simply return by checking the
`running` flag. That means of course that shutdown may take as long as the
timeout of a specific poll request. Can we tweak that? Afaik it boils down to
how long `com.amazonaws.services.kinesis.AmazonKinesis#getRecords` blocks.
--
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: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
