Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/8007#discussion_r38612546
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/cluster/YarnSchedulerBackend.scala
---
@@ -91,6 +94,36 @@ private[spark] abstract class YarnSchedulerBackend(
}
/**
+ * Override the DriverEndpoint to add extra logic for the case when an
executor is disconnected.
+ * We should check the cluster manager and find if the loss of the
executor was caused by YARN
+ * force killing it due to preemption.
+ */
+ private class YarnDriverEndpoint(rpcEnv: RpcEnv, sparkProperties:
Seq[(String, String)])
+ extends DriverEndpoint(rpcEnv, sparkProperties) {
--- End diff --
I don't see the benefit in keeping them separate. Now the developer has to
worry about which one to send a message to. If s/he accidentally picks the
wrong one then we'll just time out on a future, which is hard to debug. Also
they're not actually totally separate, since one calls into another.
Anyway I don't feel strongly about this because keeping them separate is
actually just preserving existing behavior. I just find the introduction of
another endpoint in the same file a little unnecessary.
---
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]
