Github user srowen commented on a diff in the pull request:
https://github.com/apache/spark/pull/8533#discussion_r42344504
--- Diff: core/src/main/scala/org/apache/spark/scheduler/TaskLocation.scala
---
@@ -35,16 +37,28 @@ case class ExecutorCacheTaskLocation(override val host:
String, executorId: Stri
/**
* A location on a host.
+ * The host names will be resolved to IP addresses in case
SPARK_LOCAL_HOSTNAME is unset
+ * to fix the task locality levels always being 'ANY' issue mentioned in
SPARK-10149.
*/
private [spark] case class HostTaskLocation(override val host: String)
extends TaskLocation {
- override def toString: String = host
+ override def toString: String = if ( TaskLocation.sparkLocalHostname ==
None ) {
--- End diff --
I think this is important too, I just don't know enough to evaluate the
change. BTW these should be something more like
`host.getOrElse(InetAddress.getByName(host).getHostAddress)`
I'd like to summon people like ... @andrewor14 @sryza @cmccabe. I do
suspect this will be a problem in practice for several use cases.
---
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]
