[
https://issues.apache.org/jira/browse/FLINK-4152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15385799#comment-15385799
]
ASF GitHub Bot commented on FLINK-4152:
---------------------------------------
Github user mxm commented on a diff in the pull request:
https://github.com/apache/flink/pull/2257#discussion_r71517956
--- Diff:
flink-runtime/src/main/scala/org/apache/flink/runtime/jobmanager/JobManager.scala
---
@@ -405,36 +405,23 @@ class JobManager(
currentResourceManager match {
case Some(rm) =>
- val future = (rm ? decorateMessage(new
RegisterResource(taskManager, msg)))(timeout)
- future.onComplete {
- case scala.util.Success(response) =>
- // the resource manager is available and answered
- self ! response
- case scala.util.Failure(t) =>
+ val future = (rm ? decorateMessage(new
RegisterResource(msg)))(timeout)
+ future.onFailure {
+ case t: Throwable =>
t match {
case _: TimeoutException =>
log.info("Attempt to register resource at
ResourceManager timed out. Retrying")
case _ =>
log.warn("Failure while asking ResourceManager for
RegisterResource. Retrying", t)
}
- // slow or unreachable resource manager, register anyway and
let the rm reconnect
- self ! decorateMessage(new
RegisterResourceSuccessful(taskManager, msg))
self ! decorateMessage(new ReconnectResourceManager(rm))
}(context.dispatcher)
case None =>
log.info("Task Manager Registration but not connected to
ResourceManager")
- // ResourceManager not yet available
- // sending task manager information later upon ResourceManager
registration
- self ! decorateMessage(new
RegisterResourceSuccessful(taskManager, msg))
}
- case msg: RegisterResourceSuccessful =>
-
- val originalMsg = msg.getRegistrationMessage
- val taskManager = msg.getTaskManager
-
- // ResourceManager knows about the resource, now let's try to
register TaskManager
+ // ResourceManager is told about the resource, now let's try to
register TaskManager
if (instanceManager.isRegistered(taskManager)) {
--- End diff --
Would it make sense to attempt registration of the task manager at the
JobManager first and then, if successful, notify the ResourceManager? That's
what the comment for `registeredWorkers` states now and I think it would make
sense.
> TaskManager registration exponential backoff doesn't work
> ---------------------------------------------------------
>
> Key: FLINK-4152
> URL: https://issues.apache.org/jira/browse/FLINK-4152
> Project: Flink
> Issue Type: Bug
> Components: Distributed Coordination, TaskManager, YARN Client
> Reporter: Robert Metzger
> Assignee: Till Rohrmann
> Attachments: logs.tgz
>
>
> While testing Flink 1.1 I've found that the TaskManagers are logging many
> messages when registering at the JobManager.
> This is the log file:
> https://gist.github.com/rmetzger/0cebe0419cdef4507b1e8a42e33ef294
> Its logging more than 3000 messages in less than a minute. I don't think that
> this is the expected behavior.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
