Robert Metzger created FLINK-16373:
--------------------------------------
Summary: EmbeddedLeaderService: IllegalStateException: The RPC
connection is already closed
Key: FLINK-16373
URL: https://issues.apache.org/jira/browse/FLINK-16373
Project: Flink
Issue Type: Bug
Components: Runtime / Coordination
Affects Versions: 1.11.0
Reporter: Robert Metzger
In our CI system, I see a lot of these error messages:
{code}
09:52:41,108 [flink-akka.actor.default-dispatcher-2] INFO
org.apache.flink.runtime.taskexecutor.TaskExecutor - Allocated slot
for fd36e49b9105838fc7fb8fff442ade28.
09:52:41,108 [mini-cluster-io-thread-14] WARN
org.apache.flink.runtime.highavailability.nonha.embedded.EmbeddedLeaderService
- Error notifying leader listener about new leader
java.lang.IllegalStateException: The RPC connection is already closed
at
org.apache.flink.util.Preconditions.checkState(Preconditions.java:195)
at
org.apache.flink.runtime.registration.RegisteredRpcConnection.start(RegisteredRpcConnection.java:90)
at
org.apache.flink.runtime.taskexecutor.JobLeaderService$JobManagerLeaderListener.notifyLeaderAddress(JobLeaderService.java:334)
at
org.apache.flink.runtime.highavailability.nonha.embedded.EmbeddedLeaderService$NotifyOfLeaderCall.run(EmbeddedLeaderService.java:515)
at
java.util.concurrent.CompletableFuture$AsyncRun.run(CompletableFuture.java:1640)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
09:52:41,109 [flink-akka.actor.default-dispatcher-4] INFO
org.apache.flink.runtime.taskexecutor.JobLeaderService - Resolved
JobManager address, beginning registration
{code}
Example cases
- https://transfer.sh/BAzbF/20200302.17.tar.gz
- https://transfer.sh/8344E/20200302.19.tar.gz
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
