zentol commented on a change in pull request #15553:
URL: https://github.com/apache/flink/pull/15553#discussion_r611474860
##########
File path:
flink-streaming-java/src/main/java/org/apache/flink/streaming/api/operators/collect/CollectResultFetcher.java
##########
@@ -125,7 +127,13 @@ public T next() throws IOException {
try {
response = sendRequest(buffer.getVersion(), requestOffset);
} catch (Exception e) {
- LOG.warn("An exception occurs when fetching query
results", e);
+ if (ExceptionUtils.findThrowable(
+ e,
UnavailableDispatcherOperationException.class)
Review comment:
Assuming that `collect()` is mostly used for prototyping we should cover
all cases with this change (because leader election is quick).
It would be nice if one could differentiate between timeouts at the various
layers (Dispatcher, JobManager, actual coordinator operation), but that's not
something we can fix easily :/
--
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.
For queries about this service, please contact Infrastructure at:
[email protected]
