Martijn Visser created FLINK-40906:
--------------------------------------

             Summary: DataFrameWindowITTests.test_hop_window_aggregation fails 
with a DispatcherGateway.submitJob timeout
                 Key: FLINK-40906
                 URL: https://issues.apache.org/jira/browse/FLINK-40906
             Project: Flink
          Issue Type: Bug
          Components: API / Python
    Affects Versions: 2.4.0
            Reporter: Martijn Visser


https://dev.azure.com/apache-flink/98463496-1af2-4620-8eab-a2ecc1a2e6fe/_build/results?buildId=79750&view=logs&j=00b6b22f-2e0b-5861-de1c-fd8f5bdfb924
 (master, nightly 2026-10-04, Hadoop 3.4.3 profile, JDK 17, Python 3.11)

{code}
FAILED 
pyflink/dataframe/tests/test_dataframe.py::DataFrameWindowITTests::test_hop_window_aggregation
...
E       pyflink.util.exceptions.TableException: 
org.apache.flink.table.api.TableException: Failed to execute sql
...
E       Caused by: org.apache.flink.util.FlinkException: Failed to execute job 
'collect'.
...
E       Caused by: java.util.concurrent.TimeoutException: Invocation of 
[LocalRpcInvocation(DispatcherGateway.submitJob(ExecutionPlan, Duration))] at 
recipient [pekko.tcp://flink@localhost:36770/user/rpc/dispatcher_198] timed 
out. ...
E               at 
org.apache.flink.runtime.rpc.pekko.$Proxy118.submitJob(Unknown Source)
E               at 
org.apache.flink.runtime.application.SingleJobApplication.execute(SingleJobApplication.java:94)
E               at 
org.apache.flink.runtime.dispatcher.Dispatcher.internalSubmitApplication(Dispatcher.java:1005)
...
E       Caused by: org.apache.pekko.pattern.AskTimeoutException: Ask timed out 
on [Actor[pekko://flink/user/rpc/dispatcher_198#-2025310980]] after [10000 ms]. 
Message of type [org.apache.flink.runtime.rpc.messages.LocalFencedMessage]. A 
typical reason for `AskTimeoutException` is that the recipient actor didn't 
send a reply.
{code}

The Flink log in the build's python artifact shows the dispatcher proposing 
leadership for this job at 01:48:13,316 and the 
JobMasterServiceLeadershipRunner being granted it at 01:48:44,618, after the 
submission had already timed out at 01:48:23,333. Nothing else was logged in 
that window. The other five jobs this test class submitted to the same 
dispatcher were granted leadership within 1 ms.

The test was added with FLINK-40424 and only exists on master. It passed on the 
other five python legs of the same build, and this is the only occurrence in 74 
master runs over the last seven days.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to