Github user sryza commented on a diff in the pull request:
https://github.com/apache/spark/pull/1934#discussion_r16272089
--- Diff:
yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala
---
@@ -613,29 +593,6 @@ object YarnAllocationHandler {
sparkConf)
}
- def newAllocator(
- conf: Configuration,
- amClient: AMRMClient[ContainerRequest],
- appAttemptId: ApplicationAttemptId,
- maxExecutors: Int,
- executorMemory: Int,
- executorCores: Int,
- map: collection.Map[String, collection.Set[SplitInfo]],
- sparkConf: SparkConf
- ): YarnAllocationHandler = {
- val (hostToCount, rackToCount) = generateNodeToWeight(conf, map)
--- End diff --
Previously, if no SparkContext registered, we would still attempt to create
a YarnAllocationHandler. This makes no sense, because there's nothing for us
to do with it once it's created.
This particular deletion is removing the convenience method that was used
for creating the YarnAllocationHandler in that case.
---
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]
