[
https://issues.apache.org/jira/browse/FLINK-1351?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14283932#comment-14283932
]
Till Rohrmann commented on FLINK-1351:
--------------------------------------
I could not reproduce the timeout error. What did you do exactly? Did you block
the JobManager by some breakpoints?
I also tested the behaviour in case of an unreachable job manager (wrong
address or job manager died). The TaskManager tries 10 times to connect to it
with a pause of 10 seconds in between. If he does not succeed, then he prints
an error message saying that he could not connect to the job manager and
terminates itself.
> Inconclusive error when TaskManager cannot connect to JobManager
> ----------------------------------------------------------------
>
> Key: FLINK-1351
> URL: https://issues.apache.org/jira/browse/FLINK-1351
> Project: Flink
> Issue Type: Bug
> Components: JobManager, TaskManager
> Affects Versions: 0.9
> Reporter: Stephan Ewen
> Assignee: Till Rohrmann
>
> The taskmanager currently registers at the jobmanager by resolving the akka
> URL
> {code}
> val jobManager = context.actorSelection(jobManagerAkkaURL)
> {code}
> When the actor lookup fails (actor systems cannot connect), it gives an
> unspecific timeout message. This is the case when the TaskManager cannot
> connect to the JobManager.
> This should be fixed to give a conclusive error message.
> I suggest to add a test where TaskManager is started without JobManager actor
> system being available.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
