Github user mgummelt commented on a diff in the pull request:
https://github.com/apache/spark/pull/14414#discussion_r73949486
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/MesosCoarseGrainedSchedulerBackend.scala
---
@@ -152,8 +152,13 @@ private[spark] class
MesosCoarseGrainedSchedulerBackend(
sc.sparkUser,
sc.appName,
sc.conf,
-
sc.conf.getOption("spark.mesos.driver.webui.url").orElse(sc.ui.map(_.appUIAddress))
+
sc.conf.getOption("spark.mesos.driver.webui.url").orElse(sc.ui.map(_.appUIAddress)),
+ None,
+ None,
+ sc.conf.getOption("spark.mesos.driver.frameworkId")
)
+
+ unsetFrameworkID(sc)
--- End diff --
> If no id is specified what will happen?
Mesos generates a unique framework ID when none is specified. This PR only
sets the framework ID on the first spark context created by the driver.
> What if you could use a unique id as aframeworkId plus an increment, so
the scheduler can detect all frameworks even the ones coming from multiple
contexts?
We can't guarantee no collisions with that approach. It would also be
difficult for the dispatcher to determine how many frameworks have been
started.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
