Piotr Nowojski created FLINK-7066:
-------------------------------------
Summary: Kafka integration tests failing in "airplane mode"
Key: FLINK-7066
URL: https://issues.apache.org/jira/browse/FLINK-7066
Project: Flink
Issue Type: Bug
Reporter: Piotr Nowojski
Assignee: Piotr Nowojski
Tests KafkaXXXProducerITCase are failing on my laptop in airplane mode. It
seemed to have something to do with some service listening on wrong interface,
when client tries to connect to different host. Strangely tests for Kafka010
and Kafka011 fails with different error, but there is the same fix for them
(maybe in Kafka010 original exception is masked by some other error). Kafka
0.11 tests fails like this:
{code}
35309 [flink-akka.actor.default-dispatcher-3] INFO Remoting - Starting
remoting
42445 [flink-akka.actor.default-dispatcher-3] INFO Remoting - Remoting
started; listening on addresses
:[akka.tcp://flink@fe80:0:0:0:165d:140b:f597:e019%13:54398]
42445 [main] INFO org.apache.flink.runtime.client.JobClient - Started
JobClient actor system at [fe80::165d:140b:f597:e019]:54398
42450 [flink-akka.actor.default-dispatcher-5] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Disconnect from
JobManager null.
42461 [flink-akka.actor.default-dispatcher-5] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Received
SubmitJobAndWait(JobGraph(jobId: 3b11234d116ab1ed3c1279dd73dfaab5)) but there
is no connection to a JobManager yet.
42462 [flink-akka.actor.default-dispatcher-5] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Received job
Exactly once test (3b11234d116ab1ed3c1279dd73dfaab5).
52473 [flink-akka.actor.default-dispatcher-5] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Terminate
JobClientActor.
52473 [flink-akka.actor.default-dispatcher-5] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Disconnect from
JobManager null.
org.apache.flink.runtime.client.JobExecutionException: Couldn't retrieve the
JobExecutionResult from the JobManager.
at
org.apache.flink.runtime.client.JobClient.awaitJobResult(JobClient.java:309)
...
Caused by:
org.apache.flink.runtime.client.JobClientActorConnectionTimeoutException: Lost
connection to the JobManager.
at
org.apache.flink.runtime.client.JobClientActor.handleMessage(JobClientActor.java:219)
...
{code}
I think the issue is that there is someone listening on
fe80:0:0:0:165d:140b:f597:e019 (note that this is ipv6 address from some
virtual utun0 interface on my machine), while JobClient tries to connect to
"localhost" - which fails. When I enable wifi and connect to any network and
log looks like this:
{code}
32981 [flink-akka.actor.default-dispatcher-2] INFO Remoting - Starting
remoting
32995 [flink-akka.actor.default-dispatcher-3] INFO Remoting - Remoting
started; listening on addresses :[akka.tcp://[email protected]:55576]
address = akka.tcp://[email protected]:55576
33000 [main] INFO org.apache.flink.runtime.client.JobClient - Started
JobClient actor system at 192.168.178.125:55576
33005 [flink-akka.actor.default-dispatcher-2] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Disconnect from
JobManager null.
submitJobAndWait config = {restart-strategy.fixed-delay.delay=0 s,
local.number-taskmanager=1, taskmanager.network.netty.client.numThreads=1,
metrics.reporter.my_reporter.class=org.apache.flink.metrics.jmx.JMXReporter,
jobmanager.rpc.address=localhost, taskmanager.numberOfTaskSlots=8,
taskmanager.memory.size=16, metrics.reporters=my_reporter,
taskmanager.network.netty.server.numThreads=2, jobmanager.rpc.port=55566,
query.server.enable=false}
33013 [flink-akka.actor.default-dispatcher-2] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Received
SubmitJobAndWait(JobGraph(jobId: ac67638ac85a2179a37486d507a1a008)) but there
is no connection to a JobManager yet.
33014 [flink-akka.actor.default-dispatcher-2] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Received job
Exactly once test (ac67638ac85a2179a37486d507a1a008).
33024 [flink-akka.actor.default-dispatcher-2] INFO
org.apache.flink.runtime.client.JobSubmissionClientActor - Connect to
JobManager Actor[akka.tcp://flink@localhost:55566/user/jobmanager#-1394172571].
{code}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
