[
https://issues.apache.org/jira/browse/FLINK-8163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16546392#comment-16546392
]
Chesnay Schepler commented on FLINK-8163:
-----------------------------------------
This test has some funky retrying logic that ignores most exceptions:
{code:java}
CompletableFuture<S> expected = client.getKvState(jobId, queryName, key,
keyTypeInfo, stateDescriptor);
expected.whenCompleteAsync((result, throwable) -> {
if (throwable != null) {
if (
throwable.getCause() instanceof CancellationException ||
throwable.getCause() instanceof AssertionError ||
(failForUnknownKeyOrNamespace && throwable.getCause() instanceof
UnknownKeyOrNamespaceException)
) {
resultFuture.completeExceptionally(throwable.getCause());
} else if (deadline.hasTimeLeft()) {
getKvStateIgnoringCertainExceptions(
deadline, resultFuture, client, jobId, queryName, key,
keyTypeInfo,
stateDescriptor, failForUnknownKeyOrNamespace, executor);
}
} else {
resultFuture.complete(result);
}
}, executor);{code}
When running the test locally in a loop this exception was logged on the server.
{code}
org.apache.flink.shaded.netty4.io.netty.util.internal.OutOfDirectMemoryError:
failed to allocate 16777216 byte(s) of direct memory (used: 2147483648, max:
2147483648)
{code}
The client ignores this error, infinitely retries the operation, causing the
timeout. Incidentally, on every subsequent attempt the same exception is
printed.
> NonHAQueryableStateFsBackendITCase test getting stuck on Travis
> ---------------------------------------------------------------
>
> Key: FLINK-8163
> URL: https://issues.apache.org/jira/browse/FLINK-8163
> Project: Flink
> Issue Type: Bug
> Components: Queryable State, Tests
> Affects Versions: 1.5.0
> Reporter: Till Rohrmann
> Assignee: Chesnay Schepler
> Priority: Critical
> Labels: test-stability
>
> The {{NonHAQueryableStateFsBackendITCase}} tests seems to get stuck on Travis
> producing no output for 300s.
> https://travis-ci.org/tillrohrmann/flink/jobs/307988209
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
