rkhachatryan commented on a change in pull request #17645:
URL: https://github.com/apache/flink/pull/17645#discussion_r744018286
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/channel/ChannelStateCheckpointWriter.java
##########
@@ -296,6 +298,9 @@ private void runWithChecks(RunnableWithException r) throws
Exception {
r.run();
} catch (Exception e) {
fail(e);
+ if (!findThrowable(e, IOException.class).isPresent()) {
Review comment:
Yes, my intention was to not fail the job (but only checkpoint) if the
underlying FS abstraction wraps this into RuntimeException for example (it
shouldn't).
The reasoning behind IOExceptions failing only the current checkpoint is
that they can be transient (e.g. network failure); other exceptions usually
aren't transient (e.g. programmatic errors) and therefore should fail the job.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
